CN110084426A - A kind of wind power interval prediction method based on industry and mining city - Google Patents
A kind of wind power interval prediction method based on industry and mining city Download PDFInfo
- Publication number
- CN110084426A CN110084426A CN201910341713.7A CN201910341713A CN110084426A CN 110084426 A CN110084426 A CN 110084426A CN 201910341713 A CN201910341713 A CN 201910341713A CN 110084426 A CN110084426 A CN 110084426A
- Authority
- CN
- China
- Prior art keywords
- wind power
- sample
- interval
- power
- operating condition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 65
- 238000005065 mining Methods 0.000 title claims abstract description 20
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 27
- 238000012549 training Methods 0.000 claims description 15
- 230000005611 electricity Effects 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 4
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000013461 design Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000000638 solvent extraction Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The present invention relates to wind power interval estimation technical fields, the wind power interval prediction method based on industry and mining city that specifically the present invention provides a kind of, this method completes the identification of operating condition based on fuzzy soft cluster method, and the interval prediction of wind power is completed using kernel density estimation method on this basis, the principal element for influencing wind power is determined first, wind power deterministic forecasting model is established using least square method supporting vector machine method, and clusters number is determined based on Cluster Validity, the soft cluster of fuzzy C-mean algorithm is recycled to carry out industry and mining city to history data, it is divided into multiple subintervals, the probability density function of each operating condition power prediction value and true value error is finally calculated using Density Estimator, determine the power confidence interval of each operating condition, complete interval prediction.The method of the present invention can not only predict the exact numerical of wind power, but also can predict the wind power fluctuation range under certain probability level.
Description
Technical field
The present invention relates to wind power interval estimation technical field, specifically a kind of wind power area based on industry and mining city
Between prediction technique.
Background technique
In recent years, the support energetically with country to clean energy resource, China's wind-power electricity generation cause are grown rapidly.So
And intermittence and fluctuation that wind-powered electricity generation is intrinsic, wind-electricity integration can impact the stability of power grid, it is therefore necessary to wind-powered electricity generation
Power is accurately predicted.Wind power prediction is broadly divided into deterministic forecast and interval estimation at present.Conventional wind power is pre-
It surveys and mostly uses deterministic forecast greatly, i.e., prediction result is an accurate numerical value.Compared to deterministic forecast, interval estimation is with area
Between form provide prediction result, traffic department can be given to provide more detailed wind power fluctuation information.In addition to this, exist
Wind power is predicted respectively under different operating conditions, the precision of prediction model can be greatly improved, therefore divide based on operating condition
Carrying out research to wind power interval estimation has great significance.
The method for carrying out operating condition division to power station unit at present such as usually has to calculate at frequency methods, equidensity technique, the c-means cluster
Method and k-means clustering algorithm etc., but unique foundation that these operating condition division methods only divide real hair power as operating condition, suddenly
Influence of the other factors to operating condition is omited.In addition to above-mentioned operating condition division methods, Fuzzy C-Means Cluster Algorithm (fuzzy c-
Means, FCM) due to introducing fuzzy concept, and the factor of other influences operating condition is considered, model accuracy is improved, in wind
It is widely used in terms of motor group industry and mining city.But in practical projects, transition is certainly existed between adjacent operating condition
Process, the inevitable characteristic simultaneously with adjacent operating condition of sample during this, is classified as one kind for sample merely with FCM algorithm, no
It tallies with the actual situation.For this limitation of FCM algorithm, Lv Y etc. proposes a kind of fuzzy C-mean algorithm on the basis of FCM
Soft cluster (soft fuzzy c-means, SFCM) algorithm.Sample can be divided into multiple classifications by SFCM algorithm simultaneously, from
And realize " soft cluster ".But SFCM algorithm needs to be determined in advance the class number of division, inappropriate class number necessarily drops
The validity of oligomeric class.So SFCM algorithm is used under the premise of determining preferable clustering number purpose herein, and and Density Estimator
Method, which combines, to be applied in the estimation of wind park power interval.
Summary of the invention
The wind power interval prediction method based on industry and mining city that the purpose of the present invention is to provide a kind of, it is above-mentioned to solve
The problem of being proposed in background technique.
To achieve the above object, the invention provides the following technical scheme:
A kind of wind power interval prediction method based on industry and mining city, comprising the following steps:
Stage one: modelling phase
Step1, the design data for collecting target wind farm, divide in conjunction with process of the actual motion environment to wind-power electricity generation
Analysis finds out the principal element for influencing wind power, determines the input/output variable of model；
Step2, export historical data from SCADA, and denoised, filter preprocessing, using data it is a part of as
Training sample set, a part are used as test sample collection, it is pre- to establish wind power certainty using least square method supporting vector machine method
Survey model；
Step3, it is based on Cluster Validity Index, determines clusters number, the soft clustering procedure of fuzzy C-mean algorithm is recycled to transport history
Row data carry out industry and mining city to be divided into multiple subintervals, complete the identification of operating condition；
Step4, the probability density letter that power prediction value and true value error under each operating condition are calculated using kernel density estimation method
Number, and then determine the power confidence interval of each operating condition, to complete the foundation of interval estimation model；
Stage two: wind power interval estimation stage
Each input/output variable data in Step5, online acquisition Step2, denoise data, filter preprocessing；
Step6, data pretreated in Step5 are input to the wind power interval estimation model that Step4 is obtained, obtained
Forecast interval and output to wind power.
As a further solution of the present invention: the input variable determined in the Step2 includes: local wind speed, wind direction
Angle sine value, wind angle cosine value, wheel speed, real-time torque, generator speed and previous moment reality hair power, totally 6.
As a further solution of the present invention: the output variable determined in the Step2: current time sends out power in fact,
Totally 1.
As a further solution of the present invention: in the Step2, specifically includes the following steps:
S2-1, the history data acquired from SCADA system is removed by abnormal point using Yi Lada criterion, specifically
Mathematical notation is as follows:
If sample data is x1,x2…xn, average value isDeviation isAccording to
Bessel formula calculates standard deviation:
If a certain sample data xkDeviation vk(1≤k≤n) meets | vk| 3 δ of >, then it is assumed that data are unreasonable, should
It rejects；
S2-2, by treated, history data is divided into two subsets: training subset and test subset, and sub with training
Collection completes the training of least square method supporting vector machine model, specifically includes the following steps:
Training set T by l sample at
T={ (x1,y1),(x2,y2),…,(xl,yl)}
Wherein, xi∈RnIt is input vector, yi∈RnIt is to correspond to xiOutput, enableThen least square is supported
The optimization problem of vector machine is
ω is weight vector in formula, and γ is regularization parameter, ekIt is error variance,It is from the input space to high dimensional feature
The Nonlinear Mapping in space, b are a deviators, and the Lagrange function for optimization problem is
Wherein αkIt is Lagrange multiplier, claims to correspond to αk≠ 0 sample point is supporting vector, and corresponding KKT condition is
It can be expressed as the form of following equations group
In formulaY=(y1,…,yl)T, 1=(1 ..., 1)T, α=(α1,…,αl)T,
It solves equation after obtaining α and β, for new input vector x, output valve y (x) can be calculated according to the following formula
As a further solution of the present invention: in the Step3, specifically includes the following steps:
S3-1, clustering target CHI, SSE when clusters number is respectively 2~12 are calculated, according to optimal CHI and SSE
Clustering target determines clusters number, and the calculating of CHI and SSE clustering target is as follows:
Wherein, x is the sample of subclass, XiFor subclass, K is clusters number, ciFor subclass cluster centre, N is sample
Quantity, B represent the dispersibility between class, and W represents the compactedness in class, and calculation formula is respectively as follows:
Wherein,For the average value of all samples, wk,iIndicate i-th of sample to k-th of membership from classification, i.e.,
Are as follows:
S3-2, operating condition division, the soft cluster of fuzzy C-mean algorithm are carried out to history data using the soft clustering procedure of fuzzy C-mean algorithm
Method calculates each sample to the subordinated-degree matrix U=(u of each classification first with fuzzy clustering algorithmik)i×k, then use fuzzy partitioning
Rule classifies to sample, and softening divider is then specific as follows:
If the maximum membership degree u of sample xiikMeet:
uik>0.5+0.5T-1
Then xi is uniquely divided into classification k, and T is cluster number in formula；
If the maximum membership degree u of sample xiikMeet:
uij>(T+δ)-1
Xi can belong to multiple classifications when then xi is divided into classification j, in formula, δ is degree of overlapping, be worth bigger two classifications that indicate
Lap is bigger.
As a further solution of the present invention: in the Step4, specifically includes the following steps:
S4-1, using kernel density estimation method, calculate the probability density function of power error under each operating condition, Density Estimator
Algorithm principle is as follows:
To a certain prediction error e:
E=Preal-Ppredict
The expression formula of cuclear density Multilayer networks are as follows:
Wherein, PrealFor power true value, PpredictFor predicted value, NiFor section sample number；H is bandwidth factor, is defaulted as
2；emFor error sample；K (x) is kernel function, uses gaussian kernel function here, it may be assumed that
S4-2, confidence interval most narrow under each operating condition is calculated, in the lower confidence interval given confidence level 1- α (0 < α < 1)
Meet:
P(xdown< x < xup)=1- α
Wherein, P (xdown< x < xup) indicate performance number x in section [xdown,xup] in probability, xup、xdownIt is referred to as
The upper and lower bound of confidence interval.
As a further solution of the present invention: further comprising the steps of in the Step4:
S4-3, all predicted values of traversal, obtain the envelope up and down of wind power fluctuation range, to complete wind power
Interval estimation.
Compared with prior art, it the beneficial effects of the present invention are: the present invention can be used for wind power interval estimation, and examines
Prediction result difference under different operating conditions is considered, in conjunction with Fuzzy C-Means Cluster Algorithm, has substantially increased model accuracy, simultaneously also
Have many advantages, such as that calculating is time-consuming less, applied widely.
Detailed description of the invention
Fig. 1 is the flow chart of the method for the present invention.
Fig. 2 is the deterministic forecast result based on least square method supporting vector machine.
Fig. 3 is wind power interval estimation result when confidence level is equal to 0.9.
Fig. 4 is wind power interval estimation result when confidence level is equal to 0.8.
Specific embodiment
It should be noted that in the absence of conflict, the feature in embodiment and embodiment in the present invention can phase
Mutually combination.
In the description of the present invention, it is to be understood that, term " center ", " longitudinal direction ", " transverse direction ", "upper", "lower",
The orientation or positional relationship of the instructions such as "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outside" is
It is based on the orientation or positional relationship shown in the drawings, is merely for convenience of description of the present invention and simplification of the description, rather than instruction or dark
Show that signified device or element must have a particular orientation, be constructed and operated in a specific orientation, therefore should not be understood as pair
Limitation of the invention.In addition, term " first ", " second " etc. are used for description purposes only, it is not understood to indicate or imply phase
To importance or implicitly indicate the quantity of indicated technical characteristic.The feature for defining " first ", " second " etc. as a result, can
To explicitly or implicitly include one or more of the features.In the description of the present invention, unless otherwise indicated, " multiple "
It is meant that two or more.
In the description of the present invention, it should be noted that unless otherwise clearly defined and limited, term " installation ", " phase
Even ", " connection " shall be understood in a broad sense, for example, it may be being fixedly connected, may be a detachable connection, or be integrally connected；It can
To be mechanical connection, it is also possible to be electrically connected；It can be directly connected, can also can be indirectly connected through an intermediary
Connection inside two elements.For the ordinary skill in the art, above-mentioned term can be understood by concrete condition
Concrete meaning in the present invention.
The present invention will be described in detail below with reference to the accompanying drawings and embodiments.
Embodiment one
Refering to fig. 1~4, in the embodiment of the present invention, a kind of wind power interval prediction method based on industry and mining city, including
Following steps:
Stage one: modelling phase
Step1, the design data for collecting target wind farm, divide in conjunction with process of the actual motion environment to wind-power electricity generation
Analysis finds out the principal element for influencing wind power, determines the input/output variable of model；
Step2, export historical data from SCADA, and denoised, filter preprocessing, using data it is a part of as
Training sample set, a part are used as test sample collection, it is pre- to establish wind power certainty using least square method supporting vector machine method
Survey model；
Step3, it is based on Cluster Validity Index, determines clusters number, the soft clustering procedure of fuzzy C-mean algorithm is recycled to transport history
Row data carry out industry and mining city to be divided into multiple subintervals, complete the identification of operating condition；
Step4, the probability density letter that power prediction value and true value error under each operating condition are calculated using kernel density estimation method
Number, and then determine the power confidence interval of each operating condition, to complete the foundation of interval estimation model；
Stage two: wind power interval estimation stage
Each input/output variable data in Step5, online acquisition Step2, denoise data, filter preprocessing；
Step6, data pretreated in Step5 are input to the wind power interval estimation model that Step4 is obtained, obtained
Forecast interval and output to wind power.
The input variable determined in the Step2 includes: local wind speed, wind angle sine value, wind angle cosine value, leaf
Wheel speed, real-time torque, generator speed and previous moment reality hair power, totally 6.
The output variable determined in the Step2: current time reality hair power, totally 1.
In the Step2, specifically includes the following steps:
S2-1, the history data acquired from SCADA system is removed by abnormal point using Yi Lada criterion, specifically
Mathematical notation is as follows:
If sample data is x1,x2…xn, average value isDeviation isAccording to
Bessel formula calculates standard deviation:
If a certain sample data xkDeviation vk(1≤k≤n) meets | vk| 3 δ of >, then it is assumed that data are unreasonable, should
It rejects；
S2-2, by treated, history data is divided into two subsets: training subset and test subset, and sub with training
Collection completes the training of least square method supporting vector machine model, specifically includes the following steps:
Training set T by l sample at
T={ (x1,y1),(x2,y2),…,(xl,yl)}
Wherein, xi∈RnIt is input vector, yi∈RnIt is to correspond to xiOutput, enableThen least square is supported
The optimization problem of vector machine is
ω is weight vector in formula, and γ is regularization parameter, ekIt is error variance,It is from the input space to high dimensional feature
The Nonlinear Mapping in space, b are a deviators, and the Lagrange function for optimization problem is
Wherein αkIt is Lagrange multiplier, claims to correspond to αk≠ 0 sample point is supporting vector, and corresponding KKT condition is
It can be expressed as the form of following equations group
In formulaY=(y1,…,yl)T, 1=(1 ..., 1)T, α=(α1,…,αl)T,
It solves equation after obtaining α and β, for new input vector x, output valve y (x) can be calculated according to the following formula
In the Step3, specifically includes the following steps:
S3-1, clustering target CHI, SSE when clusters number is respectively 2~12 are calculated, according to optimal CHI and SSE
Clustering target determines clusters number, and the calculating of CHI and SSE clustering target is as follows:
Wherein, x is the sample of subclass, XiFor subclass, K is clusters number, ciFor subclass cluster centre, N is sample
Quantity, B represent the dispersibility between class, and W represents the compactedness in class, and calculation formula is respectively as follows:
Wherein,For the average value of all samples, wk,iIndicate i-th of sample to k-th of membership from classification, i.e.,
Are as follows:
S3-2, operating condition division, the soft cluster of fuzzy C-mean algorithm are carried out to history data using the soft clustering procedure of fuzzy C-mean algorithm
Method calculates each sample to the subordinated-degree matrix U=(u of each classification first with fuzzy clustering algorithmik)i×k, then use fuzzy partitioning
Rule classifies to sample, and softening divider is then specific as follows:
If the maximum membership degree u of sample xiikMeet:
uik>0.5+0.5T-1
Then xi is uniquely divided into classification k, and T is cluster number in formula；
If the maximum membership degree u of sample xiikMeet:
uij>(T+δ)-1
Xi can belong to multiple classifications when then xi is divided into classification j, in formula, δ is degree of overlapping, be worth bigger two classifications that indicate
Lap is bigger.
Embodiment two
On the basis of example 1, in the Step4, specifically includes the following steps:
S4-1, using kernel density estimation method, calculate the probability density function of power error under each operating condition, Density Estimator
Algorithm principle is as follows:
To a certain prediction error e:
E=Preal-Ppredict
The expression formula of cuclear density Multilayer networks are as follows:
Wherein, PrealFor power true value, PpredictFor predicted value, NiFor section sample number；H is bandwidth factor, is defaulted as
2；emFor error sample；K (x) is kernel function, uses gaussian kernel function here, it may be assumed that
S4-2, confidence interval most narrow under each operating condition is calculated, in the lower confidence interval given confidence level 1- α (0 < α < 1)
Meet:
P(xdown< x < xup)=1- α
Wherein, P (xdown< x < xup) indicate performance number x in section [xdown,xup] in probability, xup、xdownIt is referred to as
The upper and lower bound of confidence interval.
It is further comprising the steps of in the Step4:
S4-3, all predicted values of traversal, obtain the envelope up and down of wind power fluctuation range, to complete wind power
Interval estimation.
Wind power method of interval estimation based on industry and mining city of the invention, using fuzzy C-means clustering method to going through
History data carry out operating condition division, and are all made of kernel density estimation method to each operating condition and complete wind power interval estimation, predict
Each operating condition is summarized the fluctuation model for just obtaining wind power under full working scope by the fluctuation range for having gone out each operating condition section power
It encloses, to complete interval estimation.
The present invention includes two stages: the first stage is the modelling phase, according to correlation analysis, determines the defeated of model
Enter output variable, the mathematical model of measurement object is picked out further according to fixed inputoutput data；Second stage is wind-powered electricity generation
The power prediction stage determines the fluctuation range of wind power based on obtained identification model.
Simulation result explanation
Table 1 gone out confidence level be 0.9 and 0.8 when, based on power it is isometric divide, based on FCM operating condition divide and be based on SFCM
The evaluation index for the interval estimation result that operating condition divides.δ in tablePICPFor coverage rate, coverage rate is bigger, and prediction effect is better, and
And only when coverage rate is greater than confidence level, prediction result just achieves the desired results；For interval width, guaranteeing coverage rate
Under the premise of, interval width is smaller, and prediction effect is better；δΔPFor resolution capability coefficient, value is bigger to illustrate that estimated result is better.
The evaluation index of prediction result under 1 different demarcation mode of table
By table 1 to find out, under same confidence level, the prediction result divided based on FCM operating condition will be significantly better than based on function
The prediction result that rate divides, and coverage rate, average bandwidth and the resolution capability of the interval estimation after the division of SFCM operating condition
Three indexs of coefficient are better than the prediction result of FCM operating condition division, the reason is that SFCM considers the division of boundary sample, from
The actual angle of engineering, the variation of equipment operating condition are continuously that there are corresponding transition states between different operating conditions, for mistake
The operation data under state is crossed, compared to hard cluster, it neighbouring operating condition may be distributed to simultaneously using soft cluster, more accorded with
Close the actual conditions of operating condition consecutive variations.In addition to this, since SFCM belongs to a kind of soft clustering method, more attributes have been reused
Data calculate probability density letter in later use kernel density estimation method to ensure that the sample size in each operating condition section
More close to true value when number.Under different confidence levels, evaluation index when it compared to confidence level is 0.8 that confidence level, which is 0.9,
Average bandwidth is bigger, this also complies with statistical principle.
The present invention can be used for wind power interval estimation, and consider prediction result difference under different operating conditions, in conjunction with
Fuzzy C-Means Cluster Algorithm substantially increases model accuracy, while also having many advantages, such as that calculating is time-consuming less, applied widely.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (7)
1. a kind of wind power interval prediction method based on industry and mining city, which comprises the following steps:
Stage one: modelling phase
Step1, the design data for collecting target wind farm, are analyzed in conjunction with process of the actual motion environment to wind-power electricity generation,
The principal element for influencing wind power is found out, determines the input/output variable of model；
Step2, historical data is exported from SCADA, and denoised, filter preprocessing, using a part of data as training
Sample set, a part are used as test sample collection, establish wind power deterministic forecast mould using least square method supporting vector machine method
Type；
Step3, it is based on Cluster Validity Index, determines clusters number, recycle the soft clustering procedure of fuzzy C-mean algorithm to history run number
According to industry and mining city is carried out to be divided into multiple subintervals, the identification of operating condition is completed；
Step4, the probability density function that power prediction value and true value error under each operating condition are calculated using kernel density estimation method, into
And determine the power confidence interval of each operating condition, to complete the foundation of interval estimation model；
Stage two: wind power interval estimation stage
Each input/output variable data in Step5, online acquisition Step2, denoise data, filter preprocessing；
Step6, data pretreated in Step5 are input to the wind power interval estimation model that Step4 is obtained, obtain wind
The forecast interval of electrical power and output.
2. a kind of wind power method of interval estimation based on industry and mining city according to claim 1, it is characterised in that: institute
The input variable determined in the Step2 stated includes: local wind speed, wind angle sine value, wind angle cosine value, wheel speed, reality
When torque, generator speed and previous moment hair power in fact, totally 6.
3. a kind of wind power method of interval estimation based on industry and mining city according to claim 1 or 2, feature exist
In: the output variable determined in the Step2: current time reality hair power, totally 1.
4. a kind of wind power method of interval estimation based on industry and mining city according to claim 3, it is characterised in that: institute
In the Step2 stated, specifically includes the following steps:
S2-1, the history data acquired from SCADA system is removed by abnormal point, specific mathematics using Yi Lada criterion
It is expressed as follows:
If sample data is x1,x2…xn, average value isDeviation isAccording to Bessel formula
Calculate standard deviation:
If a certain sample data xkDeviation vk(1≤k≤n) meets | vk| 3 δ of >, then it is assumed that data are unreasonable, should reject；
S2-2, by treated, history data is divided into two subsets: training subset and test subset, and complete with training subset
At the training of least square method supporting vector machine model, specifically includes the following steps:
Training set T by l sample at
T={ (x1,y1),(x2,y2),…,(xl,yl)}
Wherein, xi∈RnIt is input vector, yi∈RnIt is to correspond to xiOutput, enableThen least square method supporting vector machine
Optimization problem be
ω is weight vector in formula, and γ is regularization parameter, ekIt is error variance,It is from the input space to high-dimensional feature space
Nonlinear Mapping, b is a deviator, and the Lagrange function for optimization problem is
Wherein αkIt is Lagrange multiplier, claims to correspond to αk≠ 0 sample point is supporting vector, and corresponding KKT condition is
It can be expressed as the form of following equations group
In formulaY=(y1,…,yl)T, 1=(1 ..., 1)T, α=(α1,…,αl)T,
It solves equation after obtaining α and β, for new input vector x, output valve y (x) can be calculated according to the following formula
5. a kind of wind power method of interval estimation based on industry and mining city according to claim 4, it is characterised in that: institute
In the Step3 stated, specifically includes the following steps:
S3-1, clustering target CHI, SSE when clusters number is respectively 2~12 are calculated, is clustered according to optimal CHI and SSE
Index determines clusters number, and the calculating of CHI and SSE clustering target is as follows:
Wherein, x is the sample of subclass, XiFor subclass, K is clusters number, ciFor subclass cluster centre, N is sample number
Amount, B represent the dispersibility between class, and W represents the compactedness in class, and calculation formula is respectively as follows:
Wherein,For the average value of all samples, wk,iIndicate i-th of sample to k-th of membership from classification, i.e., are as follows:
S3-2, operating condition division is carried out to history data using the soft clustering procedure of fuzzy C-mean algorithm, the soft clustering procedure of fuzzy C-mean algorithm is first
Each sample is calculated to the subordinated-degree matrix U=(u of each classification using fuzzy clustering algorithmik)i×k, then using fuzzy partitioning rule
Classify to sample, softening divider is then specific as follows:
If the maximum membership degree u of sample xiikMeet:
uik>0.5+0.5T-1
Then xi is uniquely divided into classification k, and T is cluster number in formula；
If the maximum membership degree u of sample xiikMeet:
uij>(T+δ)-1
Xi can belong to multiple classifications when then xi is divided into classification j, and in formula, δ is degree of overlapping, be worth the bigger overlapping for indicating two classifications
Part is bigger.
6. a kind of wind power method of interval estimation based on industry and mining city according to claim 5, it is characterised in that: institute
In the Step4 stated, specifically includes the following steps:
S4-1, using kernel density estimation method, calculate the probability density function of power error under each operating condition, Density Estimator algorithm
Principle is as follows:
To a certain prediction error e:
E=Preal-Ppredict
The expression formula of cuclear density Multilayer networks are as follows:
Wherein, PrealFor power true value, PpredictFor predicted value, NiFor section sample number；H is bandwidth factor, is defaulted as 2；em
For error sample；K (x) is kernel function, uses gaussian kernel function here, it may be assumed that
S4-2, confidence interval most narrow under each operating condition is calculated, met in the lower confidence interval (0 < α < 1) given confidence level 1- α:
P(xdown< x < xup)=1- α
Wherein, P (xdown< x < xup) indicate performance number x in section [xdown,xup] in probability, xup、xdownIt is referred to as confidence
The upper and lower bound in section.
7. a kind of wind power method of interval estimation based on industry and mining city according to claim 6, it is characterised in that: institute
It is further comprising the steps of in the Step4 stated:
S4-3, all predicted values of traversal, obtain the envelope up and down of wind power fluctuation range, to complete the area of wind power
Between estimate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910341713.7A CN110084426A (en) | 2019-04-26 | 2019-04-26 | A kind of wind power interval prediction method based on industry and mining city |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910341713.7A CN110084426A (en) | 2019-04-26 | 2019-04-26 | A kind of wind power interval prediction method based on industry and mining city |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110084426A true CN110084426A (en) | 2019-08-02 |
Family
ID=67416793
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910341713.7A Pending CN110084426A (en) | 2019-04-26 | 2019-04-26 | A kind of wind power interval prediction method based on industry and mining city |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110084426A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110659672A (en) * | 2019-09-02 | 2020-01-07 | 国电新能源技术研究院有限公司 | Wind turbine generator output step uncertainty prediction method and device |
| CN112633630A (en) * | 2020-11-23 | 2021-04-09 | 贵州电网有限责任公司 | Multi-energy power fluctuation interval identification method |
| CN113468811A (en) * | 2021-07-06 | 2021-10-01 | 国网陕西省电力公司 | Power grid reserve capacity probabilistic dynamic evaluation method, system, terminal and readable storage medium containing new energy unit |
-
2019
- 2019-04-26 CN CN201910341713.7A patent/CN110084426A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110659672A (en) * | 2019-09-02 | 2020-01-07 | 国电新能源技术研究院有限公司 | Wind turbine generator output step uncertainty prediction method and device |
| CN110659672B (en) * | 2019-09-02 | 2023-09-26 | 国电新能源技术研究院有限公司 | Method and device for predicting step-by-step uncertainty of output of wind turbine generator |
| CN112633630A (en) * | 2020-11-23 | 2021-04-09 | 贵州电网有限责任公司 | Multi-energy power fluctuation interval identification method |
| CN113468811A (en) * | 2021-07-06 | 2021-10-01 | 国网陕西省电力公司 | Power grid reserve capacity probabilistic dynamic evaluation method, system, terminal and readable storage medium containing new energy unit |
| CN113468811B (en) * | 2021-07-06 | 2024-03-08 | 国网陕西省电力公司 | Power grid reserve capacity probabilistic dynamic assessment method and system containing new energy unit |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Han et al. | A PV power interval forecasting based on seasonal model and nonparametric estimation algorithm | |
| CN110084426A (en) | A kind of wind power interval prediction method based on industry and mining city | |
| CN108985380B (en) | Point switch fault identification method based on cluster integration | |
| CN102236795B (en) | Method for forecasting wind speed in wind power station | |
| CN103488869A (en) | Wind power generation short-term load forecast method of least squares support vector machine | |
| Niu et al. | Study of degradation of fuel cell stack based on the collected high-dimensional data and clustering algorithms calculations | |
| CN107038167A (en) | Big data excavating analysis system and its analysis method based on model evaluation | |
| CN104933483A (en) | Wind power forecasting method dividing based on weather process | |
| CN110503256A (en) | Short-term load forecasting method and system based on big data technology | |
| CN104112062B (en) | The acquisition methods of wind-resources distribution based on interpolation method | |
| CN114676742A (en) | Power grid abnormal electricity utilization detection method based on attention mechanism and residual error network | |
| CN109902339A (en) | A kind of Fault Diagnosis of Roller Bearings based on IAGA-SVM | |
| CN112791997A (en) | Method for cascade utilization and screening of retired battery | |
| CN109934456A (en) | A kind of method and system for acquisition operational system progress intelligent trouble detection | |
| CN103425994A (en) | Feature selecting method for pattern classification | |
| CN106712065A (en) | Prediction method of indexes of influence of electric card charging on power distribution network active loss | |
| CN112288157A (en) | Wind power plant power prediction method based on fuzzy clustering and deep reinforcement learning | |
| CN116050867A (en) | Data-driven power system transient stability evaluation method with interpretability | |
| CN115169659A (en) | Ultra-short-term roof photovoltaic power prediction method and system based on multiple classifiers | |
| CN113392497B (en) | Method for measuring field performance degradation of photovoltaic module according to geographical region | |
| CN106980874A (en) | A kind of multi-time Scales dimension data fusion method towards distribution big data | |
| CN109902743A (en) | A kind of Wind turbines output power predicting method | |
| CN111806443A (en) | Pure electric heavy truck unmanned energy consumption optimization method | |
| CN109389309A (en) | A kind of wind-driven generator performance estimating method, system, equipment and computer media | |
| CN112561322B (en) | Charging pile site selection method based on machine learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190802 |
|
| WD01 | Invention patent application deemed withdrawn after publication |