CN105888987A - Wind generating set performance assessment method based on correlation analysis - Google Patents

Abstract

The invention relates to a wind generating set performance assessment method based on the correlation analysis. The method comprises the following steps that firstly, data are collected through a big-data platform; secondly, a generating set to be analyzed is determined; thirdly, the set in which correlation needs to be determined is subjected to the correlation analysis through different unit wind speeds with time as the sequence; fourthly, a correction unit is selected, and the relative power generating amounts of the correction unit are obtained; and fifthly, the average value of the relative power generating amounts is solved, when the relative power generating amount of the generating set to be analyzed is lower than the average power generating amount (k*100)%, the performance of the generating set to be analyzed is not good enough, and the generating set to be analyzed is in urgent need of rectification, wherein k is a constant. According to the wind generating set performance assessment method, a multi-level correlation and consistency analysis method is utilized, the related unit is determined, the influence of complex wind conditions on the performance assessment is eliminated, the relative power generating amount of the related unit is directly applied to serve as the assessment standard, the influence of good and bad wind resources on power generating is eliminated, and the assessment result is more reliable.

Description

Wind power generating set performance estimating method based on correlation analysis

Technical field

The present invention relates to a kind of wind power generating set performance estimating method, particularly a kind of based on correlation analysis Wind power generating set performance estimating method.

Background technology

Wind-power electricity generation is fast-developing in recent years and gets the attention and supports.In " 12 " planning clearly Proposing, by 2015, installed capacity of wind-driven power reached 1.04 hundred million kilowatts, rose to this numeral of the year two thousand twenty 200000000 kilowatts.See with practical conditions, adding new capacity 2335.05 ten thousand kilowatts in 2014, increased than 2013 Long 45.1%, accumulative installed capacity 1.1476339 hundred million kilowatts, than newly-increased 25.5.% in 2013.Wind in 2014 Electricity generated energy rough estimates are more than 1,500 hundred million kilowatt hours, but before grid-connected installed capacity reaches expection standard Put and do not reach intended 180,000,000,000 kilowatt hours.

The height of wind-power electricity generation electric field electricity-generating amount depends on the performance quality of wind power generating set, and it is also wind field The key that profitability is strong and weak.Accurately and reliably assessment wind power generating set performance can help wind energy turbine set to formulate O&M strategy, on purpose arranges maintenance.Although wind power generating set performance quality intuitively depends on it The height of generated energy, but during actual assessment, there is wind regime between wind power generating set and differ and lack All difficult problems such as weary contrast means, wind generating set structure and control system are complicated, are difficult to single dependence The height of wind turbine power generation amount goes to assess its performance, and the most one well can not consider many The mode of kind of factor goes to assess wind power generating set, and this makes to analyze not good enough former of wind-power electricity generation generated energy Because of so rectified and improved.Accordingly, it is desirable to provide a kind of method by data statistics, analysis goes assessment difference The method that between generating set, performance is good and bad, and provide respective algorithms to find the wind power generating set of poor-performing, High-quality operation, generated energy for wind field promote offer reference.

Summary of the invention

The technical problem to be solved be to provide a kind of can comprehensive considering various effects based on dependency The wind power generating set performance estimating method analyzed.

It is as follows that the present invention solves the technical scheme that above-mentioned technical problem used:

Wind power generating set performance estimating method based on correlation analysis, comprises the following steps:

S1. by big data platform collect data, data be the wind power generating set of different model with the time be The wind speed of sequence, power, actual power generation, turbulence intensity, wind direction and mean wind speed；

S2. determining generating set to be analyzed, extracting all models from big data platform is generating set to be analyzed Wind speed, power, actual power generation, turbulence intensity, wind direction and mean wind speed；

S3. utilize the different unit wind speed with the time as sequence to it needs to be determined that the unit of dependency carries out dependency Analyze:

S31. the generating set to be analyzed wind series with the time as sequence is set as x_i, it is thus necessary to determine that relevant The unit of the property wind series with the time as sequence is as y_i, coefficient R the most between the two₁For:

$R_1=\frac{\mathrm{\Σ}(x_i-\stackrel{\‾}{x})(y_i-\stackrel{\‾}{y})}{\sqrt{\mathrm{\Σ}{(x_i-\stackrel{\‾}{x})}^2{(y_i-\stackrel{\‾}{y})}^2}}$

S32.R₁Value the highest, represent that the dependency of the two is the highest, whenTime represent the two be correlated with, For correlative relationship unit, wherein,

S33. repeat S31 and S32, obtain all it needs to be determined that unit and the generating to be analyzed of dependency The correlation coefficient of unit, and correlative relationship unit；

S4. choose correction unit, and try to achieve the relative generated energy revising unit:

S41. from dependency unit, prevailing wind direction is chosen consistent, and whole wind direction frequency distribution correlation coefficientUnit, wherein,

Wind direction frequency distribution coefficient R₂Computational methods as follows:

$R_2=\frac{\mathrm{\Σ}\left(f_1\right(i)-\stackrel{\‾}{f_2\left(i\right)})\left(f_2\right(i)-\stackrel{\‾}{f_2\left(i\right)})}{\sqrt{\mathrm{\Σ}{\left(f_1\right(i)-\stackrel{\‾}{f_1\left(i\right)})}^2{\left(f_2\right(i)-\stackrel{\‾}{f_2\left(i\right)})}^2}}$

Wherein, i is the sequence number that wind direction is interval, i=1,2,3 ... and, f (i) is that unit falls in i interval Wind direction probability；

S42. the unit that reselection turbulence intensity is close from the unit that S41 screens, screening technique As follows:

Turbulence intensity interval is (C-c, C+c), wherein turbulence intensity centered by C, and c is interval range, when from The unit that S41 screens and the generating set to be analyzed turbulent flow within the same time period fall in identical interval Time, then the two turbulent flow is close；

S43. calculating relative generated energy, method is as follows:

Wherein, W_ActualFor the actual power generation of generating set to be analyzed, W_RelativelyPhase for generating set to be analyzed To generated energy,Mean wind speed for the unit that S42 screens；

S5. the meansigma methods of relative generated energy is sought, when generating set to be analyzed relative to generated energy less than average generating During amount (k*100) %, generating set poor performance the most to be analyzed, it is badly in need of rectification, wherein k is constant.

The present invention utilizes multi-level dependency, consistency analysis method, it is determined that relevant unit, eliminates The complicated wind regime impact on Performance Evaluation, directly the relative generated energy of the relevant unit of application is as evaluation criteria, Eliminating the good and bad impact on generating of wind-resources, assessment result is more reliable.

As preferably, in S1, the data collected by big data platform, time interval is consistent, and time model Enclose consistent.

As preferably, in S41, the consistent determination method of prevailing wind direction is as follows:

Wind direction interval is (A-a, A+a), wherein wind angle centered by A, and a is interval range, generating to be analyzed Unit falls, and to account for the ratio of conceptual data be P to the wind direction data in wind direction interval₁, the unit of relation to be determined is corresponding Ratio be P₂, the absolute value of the two difference is PP:

PP=| P₁-P₂|

When PP≤5%, the two prevailing wind direction is consistent.

As preferably, in S41, A is generating set to be analyzed with 1 ° for the highest angle of wind frequency division cloth during interval Degree, a is 5 °.

As preferably, wind direction interval is, in the wind direction data of a year, compares the interval of equal size, and Fall ratio in wind direction interval of the wind direction of all units is maximum.

As preferably, in S41, wind direction interval is (B (i)-b, B (i)+b), and wherein i is the sequence number that wind direction is interval, I=1,2,3 ..., B (i) is the center wind direction in this interval, and b is interval range, phase sized by i interval Together, mutually do not occur simultaneously and cover 0-360 ° of scope.

As preferably, the value of B and b is, B=0.25+0.5i, b=0.25.

As preferably, in S5, the computational methods of the meansigma methods of relative generated energy are as follows:

Wherein, n is the generating set quantity after S4 screens, and k is constant, and 0.8≤k ＜ 1, when treating point Analysis unitTime, unit performance the most to be analyzed is not good enough.K value is the biggest, and the assessment to unit is got over Strictly.

The present invention compared with the existing technology has the following advantages and effect:

1, big data platform is utilized to collect abundant data due to the present invention so that assessment result is more reliable, Closer to reality.

2, multi-level dependency, consistency analysis method are utilized due to the present invention so that assessment result more may be used Lean on, closer to reality.

3, it is determined by relevant unit due to the present invention, eliminates the complicated wind regime impact on Performance Evaluation, directly Scoop out the relative generated energy with relevant unit as evaluation criteria, eliminate the good and bad impact on generating of wind-resources, Assessment result is made to have higher accuracy.

4, due to the present invention use the generated energy of big data as evaluation criteria, using the labyrinth of blower fan as Operation black box processes, and simplifies estimation flow, more engineering significance.

Detailed description of the invention

Below in conjunction with embodiment, the present invention is described in further detail, and following example are to the present invention Explain and the invention is not limited in following example.

Embodiment 1:

In the present embodiment, wind power generating set performance estimating method based on correlation analysis, including following Step:

S1. by big data platform collect data, data be the wind power generating set of different model with the time be The wind speed of sequence, power, actual power generation, turbulence intensity, wind direction and mean wind speed；

S2. determining generating set to be analyzed, extracting all models from big data platform is generating set to be analyzed Wind speed, power, actual power generation, turbulence intensity, wind direction and mean wind speed；

S3. utilize the different unit wind speed with the time as sequence to it needs to be determined that the unit of dependency carries out dependency Analyze:

S31. the generating set to be analyzed wind series with the time as sequence is set as x_i, it is thus necessary to determine that relevant The unit of the property wind series with the time as sequence is as y_i, coefficient R the most between the two₁For:

$R_1=\frac{\mathrm{\Σ}(x_i-\stackrel{\‾}{x})(y_i-\stackrel{\‾}{y})}{\sqrt{\mathrm{\Σ}{(x_i-\stackrel{\‾}{x})}^2{(y_i-\stackrel{\‾}{y})}^2}}$

S32.R₁Value the highest, represent that the dependency of the two is the highest, whenTime represent the two be correlated with, For correlative relationship unit, wherein,

S33. repeat S31 and S32, obtain all it needs to be determined that unit and the generating to be analyzed of dependency The correlation coefficient of unit, and correlative relationship unit；

S4. choose correction unit, and try to achieve the relative generated energy revising unit:

S41. from dependency unit, prevailing wind direction is chosen consistent, and whole wind direction frequency distribution correlation coefficientUnit, wherein,

Wind direction frequency distribution coefficient R₂Computational methods as follows:

$R_2=\frac{\mathrm{\Σ}\left(f_1\right(i)-\stackrel{\‾}{f_2\left(i\right)})\left(f_2\right(i)-\stackrel{\‾}{f_2\left(i\right)})}{\sqrt{\mathrm{\Σ}{\left(f_1\right(i)-\stackrel{\‾}{f_1\left(i\right)})}^2{\left(f_2\right(i)-\stackrel{\‾}{f_2\left(i\right)})}^2}}$

Wherein, i is the sequence number that wind direction is interval, i=1,2,3 ... and, f (i) is that unit falls in i interval Wind direction probability；

S42. the unit that reselection turbulence intensity is close from the unit that S41 screens, screening technique As follows:

Turbulence intensity interval is (C-c, C+c), wherein turbulence intensity centered by C, and c is interval range, when from The unit that S41 screens and the generating set to be analyzed turbulent flow within the same time period fall in identical interval Time, then the two turbulent flow is close；

S43. calculating relative generated energy, method is as follows:

Wherein, W_ActualFor the actual power generation of generating set to be analyzed, W_RelativelyPhase for generating set to be analyzed To generated energy,Mean wind speed for the unit that S42 screens；

S5. the meansigma methods of relative generated energy is sought, when generating set to be analyzed relative to generated energy less than average generating During amount (k*100) %, generating set poor performance the most to be analyzed, it is badly in need of rectification, wherein k is constant.

In above-mentioned S1, collecting mainly for process below provides enough data supportings of big data.

In above-mentioned S1, the data collected by big data platform, time interval is consistent, and time range one Cause.

In above-mentioned S2, the purpose of the generating set choosing same model be ensure it include impeller diameter, Hub height, control system, drive system, electromotor, sensor-based system etc. are to use identical equipment and skill Art.

In above-mentioned S41, the consistent determination method of prevailing wind direction is as follows:

Wind direction interval is (A-a, A+a), wherein wind angle centered by A, and a is interval range, generating to be analyzed Unit falls, and to account for the ratio of conceptual data be P to the wind direction data in wind direction interval₁, the unit of relation to be determined is corresponding Ratio be P₂, the absolute value of the two difference is PP:

PP=| P₁-P₂|

When PP≤5%, the two prevailing wind direction is consistent.

In above-mentioned S41, A is generating set to be analyzed with 1 ° for the highest angle of wind frequency division cloth during interval, A is 5 °.

Above-mentioned wind direction interval is, in the wind direction data of a year, compares the interval of equal size, and all Fall ratio in wind direction interval of the wind direction of unit is maximum.

In above-mentioned S41, wind direction interval is (B (i)-b, B (i)+b), and wherein i is the sequence number that wind direction is interval, i=1, 2,3 ..., B (i) is the center wind direction in this interval, and b is interval range, identical sized by i interval, phase Do not occur simultaneously and cover 0-360 ° of scope mutually.

The value of above-mentioned B and b is, B=0.25+0.5i, b=0.25.

In above-mentioned S5, the computational methods of the meansigma methods of relative generated energy are as follows:

Wherein, n is the generating set quantity after S4 screens, and k is constant, and 0.8≤k ＜ 1, when treating point Analysis unitTime, unit performance the most to be analyzed is not good enough.Require according to different assessment or Different requirements to unit, can determine the concrete numerical value of k according to actual needs.

Embodiment 2:

The present embodiment is similar to Example 1, and it is different only in that:

S5. the meansigma methods of relative generated energy is sought, when generating set to be analyzed relative to generated energy less than average generating When measuring 80%, generating set poor performance the most to be analyzed, it is badly in need of rectification, wherein k is constant.

In step s 5, the computational methods of the meansigma methods of relative generated energy are as follows:

Wherein, n is the generating set quantity after S4 screens, and takes k=0.8 in the present embodiment, when treating point Analysis unitTime, unit performance the most to be analyzed is not good enough.

The value of k is the biggest, and the assessment to unit is the strictest, therefore when k takes 0.8, to the assessment of unit relatively For loosely, the most strictly.

Embodiment 3:

The present embodiment is similar to embodiment 1 and/or embodiment 2, and it is different only in that:

S5. the meansigma methods of relative generated energy is sought, when generating set to be analyzed relative to generated energy less than average generating When measuring 90%, generating set poor performance the most to be analyzed, it is badly in need of rectification, wherein k is constant.

In step s 5, the computational methods of the meansigma methods of relative generated energy are as follows:

Wherein, n is the generating set quantity after S4 screens, and takes k=0.9 in the present embodiment, when treating point Analysis unitTime, unit performance the most to be analyzed is not good enough.

The value of k is the biggest, and the assessment to unit is the strictest, therefore when k takes 0.9, to the assessment of unit relatively For strictly.

Furthermore, it is necessary to illustrate, the specific embodiment described in this specification, the shape of its parts and components Shape, be named title etc. can be different.All done according to structure, feature and the principle described in inventional idea of the present invention Equivalence or simple change, be all included in the protection domain of patent of the present invention.The technical field of the invention Technical staff described specific embodiment can be made various amendment or supplement or use similar Mode substitutes, without departing from the structure of the present invention or surmount scope defined in the claims, all Protection scope of the present invention should be belonged to.

Claims (8)

1. a wind power generating set performance estimating method based on correlation analysis, is characterized in that: include with Lower step:

S1. by big data platform collect data, described data be the wind power generating set of different model with time Between be the wind speed of sequence, power, actual power generation, turbulence intensity, wind direction and mean wind speed；

S2. determining generating set to be analyzed, extracting all models from big data platform is generating set to be analyzed Wind speed, power, actual power generation, turbulence intensity, wind direction and mean wind speed；

S3. utilize the different unit wind speed with the time as sequence to it needs to be determined that the unit of dependency carries out dependency Analyze:

S31. the generating set to be analyzed wind series with the time as sequence is set as x_i, it is thus necessary to determine that relevant The unit of the property wind series with the time as sequence is as y_i, coefficient R the most between the two₁For:

$R_1=\frac{\mathrm{\Σ}(x_i-\stackrel{\‾}{x})(y_i-\stackrel{\‾}{y})}{\sqrt{\mathrm{\Σ}{(x_i-\stackrel{\‾}{x})}^2{(y_i-\stackrel{\‾}{y})}^2}}$

S32.R₁Value the highest, represent that the dependency of the two is the highest, whenTime represent the two be correlated with, For correlative relationship unit, wherein,

S33. repeat S31 and S32, obtain all it needs to be determined that unit and the generating to be analyzed of dependency The correlation coefficient of unit, and correlative relationship unit；

S4. choose correction unit, and try to achieve the relative generated energy revising unit:

S41. from dependency unit, prevailing wind direction is chosen consistent, and whole wind direction frequency distribution correlation coefficientUnit, wherein,

Wind direction frequency distribution coefficient R₂Computational methods as follows:

$R_2=\frac{\mathrm{\Σ}\left(f_1\right(i)-\stackrel{\‾}{f_2\left(i\right)})\left(f_2\right(i)-\stackrel{\‾}{f_2\left(i\right)})}{\sqrt{\mathrm{\Σ}{\left(f_1\right(i)-\stackrel{\‾}{f_1\left(i\right)})}^2{\left(f_2\right(i)-\stackrel{\‾}{f_2\left(i\right)})}^2}}$

Wherein, i is the sequence number that wind direction is interval, i=1,2,3 ... and, f (i) is that unit falls in i interval Wind direction probability；

S42. the unit that reselection turbulence intensity is close from the unit that S41 screens, screening technique As follows:

Turbulence intensity interval is (C-c, C+c), wherein turbulence intensity centered by C, and c is interval range, when from The unit that S41 screens and the generating set to be analyzed turbulent flow within the same time period fall in identical interval Time, then the two turbulent flow is close；

S43. calculating relative generated energy, method is as follows:

Wherein, W_ActualFor the actual power generation of generating set to be analyzed, W_RelativelyPhase for generating set to be analyzed To generated energy,Mean wind speed for the unit that S42 screens；

S5. the meansigma methods of relative generated energy is sought, when generating set to be analyzed relative to generated energy less than average generating During amount (k*100) %, generating set poor performance the most to be analyzed, it is badly in need of rectification, wherein k is constant.

Wind power generating set performance estimating method based on correlation analysis the most according to claim 1, It is characterized in that: in described S1, the data collected by big data platform, time interval is consistent, and the time Scope is consistent.

Wind power generating set performance estimating method based on correlation analysis the most according to claim 1, It is characterized in that: in described S41, the consistent determination method of prevailing wind direction is as follows:

Wind direction interval is (A-a, A+a), wherein wind angle centered by A, and a is interval range, generating to be analyzed Unit falls, and to account for the ratio of conceptual data be P to the wind direction data in described wind direction interval₁, the unit of relation to be determined Corresponding ratio is P₂, the absolute value of the two difference is PP:

PP=| P₁-P₂|

When PP≤5%, the two prevailing wind direction is consistent.

Wind power generating set performance estimating method based on correlation analysis the most according to claim 3, It is characterized in that: in described S41, A is that generating set to be analyzed is the highest with 1 ° for wind frequency division cloth during interval Angle, a is 5 °.

Wind power generating set performance estimating method based on correlation analysis the most according to claim 3, It is characterized in that: described wind direction interval is, in the wind direction data of a year, compare the interval of equal size, And the wind direction of all units ratio in described wind direction interval that falls is maximum.

Wind power generating set performance estimating method based on correlation analysis the most according to claim 1, It is characterized in that: in described S41, wind direction interval is (B (i)-b, B (i)+b), and wherein i is the sequence that wind direction is interval Number, i=1,2,3 ..., B (i) is the center wind direction in this interval, and b is interval range, and i interval is big Little identical, mutually do not occur simultaneously and cover 0-360 ° of scope.

Wind power generating set performance estimating method based on correlation analysis the most according to claim 6, It is characterized in that: the value of described B and b is, B=0.25+0.5i, b=0.25.

Wind power generating set performance estimating method based on correlation analysis the most according to claim 1, It is characterized in that: in described S5, the computational methods of the meansigma methods of relative generated energy are as follows:

Wherein, n is the generating set quantity after S4 screens, and k is constant, and 0.8≤k ＜ 1, when treating point Analysis unitTime, unit performance the most to be analyzed is not good enough.