CN104200290B - Wind power forecast method - Google Patents
Wind power forecast method Download PDFInfo
- Publication number
- CN104200290B CN104200290B CN201410499648.8A CN201410499648A CN104200290B CN 104200290 B CN104200290 B CN 104200290B CN 201410499648 A CN201410499648 A CN 201410499648A CN 104200290 B CN104200290 B CN 104200290B
- Authority
- CN
- China
- Prior art keywords
- wind power
- record
- meteorological factor
- meteorological
- history
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 14
- 238000006116 polymerization reaction Methods 0.000 claims abstract description 19
- 238000013507 mapping Methods 0.000 claims abstract description 11
- 239000011159 matrix material Substances 0.000 claims description 21
- 238000013277 forecasting method Methods 0.000 claims description 10
- 230000035945 sensitivity Effects 0.000 abstract description 2
- 238000007796 conventional method Methods 0.000 abstract 1
- 241001269238 Data Species 0.000 description 1
- 230000003399 chemotactic effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 239000012467 final product Substances 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a wind power forecast method. The method is realized through the way that: taking time points of wind power forecasting as basic points; on the basis of standardization of historical data, taking meteorology of the basic points as core elements to perform proximity polymerization on historical meteorological records; performing decorrelation dimension reduction on the meteorological factors of the meteorological records in the polymerization to obtain independent factors; then building a function relationship from independent factor values to standardized values of the wind power according to the mapping relation of the regularization radial basis function network to realize wind power forecast. By adopting the wind power forecast method provided by the invention, not only can problems of high nonlinearity degree of the 'meteorological factors-wind power' mapping relation and large calculated amount, caused by complexity of the meteorological factors, of the conventional methods be solved, but also the problem of poor sensitivity of the conventional wind power forecast methods on fluctuation modes and ranges of the meteorological factors can be solved, so that the wind power forecast accuracy is improved and a high forecast speed is realized.
Description
Technical field
The invention belongs to electric power project engineering field, and in particular to a kind of wind power forecasting method.
Background technology
At present, wind-power electricity generation is in global fast development., due to intrinsic undulatory property and intermittence, it is by electric power for wind power
The sufficient degree that net receives depends on the accuracy and rapidity of its Forecasting Methodology.
Existing wind power forecasting method, with multiple time points even a day as a cycle, set up in the cycle from
Meteorological factor to wind power a mapping relations, realize wind power prediction.Once set up this mapping relations, so that it may
To predict the wind power of each time point in a cycle.Existing method predetermined speed is fast, but due to consider in a cycle
Time point is more, Meteorology Factor Change complicated, mapping relations not only nonlinearity it is high, to meteorological factor fluctuation model and amplitude not
Enough sensitivities, accuracy is not high, and the foundation of mapping relations needs a large amount of historical datas, computationally intensive.
The content of the invention
In order to solve the defect of existing method, it is an object of the invention to provide a kind of wind power forecasting method, to improve
The accuracy of wind power prediction, and reduce amount of calculation.
The present invention is to provide a kind of wind power forecasting method of basic point meteorology neighbour's polymerization.In the present invention, with pre-
The time point for surveying wind power is basic point, is core to history with the meteorology of basic point on the basis of standardized to historical data
Meteorological record carries out neighbour's polymerization.Again to the meteorological factor decorrelation dimensionality reduction of meteorological record in polymerization, obtain independent factor.Then
According to the mapping relations of regularization radial primary function network independent factor value is set up to the functional relationship of wind power standardized value,
Realize wind power prediction.
Specifically, the wind power forecasting method that the present invention is provided, comprises the following steps:
(1) to predict the time point of wind power as basic point, collection affects the historical data of the meteorological factor of wind power
With the matrix form of the historical data of wind power, history of forming meteorological factor record and history wind power record;
(2) basic point meteorological factor record, history meteorological factor record and history wind power record are carried out into respectively standard
Change;
(3) standardization result with basic point meteorological factor record as core, tie by the standardization that history meteorological factor is recorded
Fruit filters out from small to large n record by Euclidean distance, obtains the matrix form of the meteorological neighbour's polymerization of basic point, and wherein n is for just
Integer;
(4) to the meteorological factor record decorrelation dimensionality reduction in neighbour's polymerization, independent factor value matrix is obtained;
(5) according to the mapping relations of regularization radial primary function network independent factor value is set up to wind power standardized value
Functional relationship;
(6) prediction of basic point wind power is carried out by the functional relationship.
Wind speed, wind direction, temperature, 5 meteorological factor of air pressure and humidity are chosen in the step (1).
N values are in the step (3):5≤n≤10.
The inventive method is not only solved in existing method caused by meteorological factor complexity " meteorological factor-wind power "
The high and computationally intensive problem of mapping relations nonlinearity, and existing wind power forecasting method is solved to meteorological factor ripple
Dynamic model formula and insufficient amplitude sensitive issue, improve the accuracy of wind power prediction, and predetermined speed is fast.
Description of the drawings
Accompanying drawing 1 is flowchart of the invention.
Specific embodiment
As shown in figure 1, to realize the inventive method flow chart, illustrate below by way of a specific embodiment, this
Bright method is comprised the following steps:
(1) choose " wind speed, wind direction, temperature, air pressure, humidity " 5 meteorological factor according to the generation process of wind power to make
For the influence factor of wind power.This 5 meteorological factor are indicated successively using subscript numbering i=1,2 ..., 5.Basic point and history
Meteorological factor record matrix form be:
Wherein:uiT () is record of the meteorological factor in moment t that numbering is i, is known quantity;Column vector uiBe numbering be i
Meteorological factor all records;T=0 and t=1,2 ..., k is respectively that basic point and history meteorological factor record the corresponding moment
Backward numbering, k be history meteorological factor record number, be known quantity.That is t=0 is the time point for predicting wind power, is gone through
The general record chosen in 3~7 days before basic point of history meteorological factor record, typically k point was uniformly taken in 3~7 days, and k is general
Value is 288~672.
History wind power record column vector form be:
P=[p (1), p (2) ... p (t) ..., p (k)]T
Wherein:P (t) be wind power historical juncture t record, be known quantity;Column vector p is all of wind power
Historical record;Superscript notation T is the matrix transpose operator in algebra.
(2) following two formulas standardization are pressed respectively to meteorological factor record and history wind power record:
v′i(t)=[ui(t)-ρ(ui)]/σ(ui)
Y (t)=[p (t)-ρ (p)]/σ (p)
Wherein:v′iT () is meteorological factor record uiThe standardization result of (t), ρ (ui) and σ (ui) it is respectively column vector uiIn
The statistics average of element and standard deviation;Y (t) is the standardization result of history wind power record p (t), and ρ (p) and σ (p) divides
It is not the statistics average and standard deviation of element in column vector p.After standardization, the different meteorological factor of 5 individual character quality and quantity guiding principles is big
The same chemotactic of little comparable, power of influence.
(3) standardization result [v ' recorded with basic point meteorological factor1(0), v '2(0) ..., v 'i..., v ' (0)5(0)]
For core, to t=1,2 ..., the standardization result of k history meteorological factor record press Euclidean distanceThe individual records of n (5≤n≤10), referred to as basic point meteorology neighbour polymerization, letter are filtered out from small to large
Claim neighbour polymerization.To respectively record in neighbour's polymerization the corresponding moment renumber in reverse order, still with t (t=1,2 ..., n) table
Show these numberings, obtain the matrix form recorded in neighbour's polymerization:
Wherein:viT () is record of the meteorological factor in moment t that numbering is i in neighbour's polymerization;Column vector viIt is that neighbour gathers
Numbering is all records of the meteorological factor of i in conjunction.
(4) decorrelation dimensionality reduction as the following formula is recorded to the meteorological factor in neighbour's polymerization, obtains the rectangular of independent factor value
Formula:[x1, x2..., xh..., xm]=[v1, v2..., vi..., v5][a1, a2..., ah..., am]
Wherein:Subscript h (h=1,2 ..., m) be independent factor numbering;Column vector xhBe numbering be h independent factor
In value not in the same time;Column vector ahIt is matrix [v1, v2..., vi..., v5] the big characteristic roots pair of row correlation matrix h
The unit character vector answered, it is total that m is that the first big characteristic root of the row correlation matrix accounts for characteristic root to the big characteristic root sums of m
The ratio of sum is more than 0.85 positive integer for determining, is referred to as the number of independent factor.
(5) according to the mapping relations of regularization radial primary function network independent factor value is set up to wind power standardized value
Functional relationship:
Wherein:Y (t) be historical juncture t (t=1,2 ..., n) the standardized value of wind power;Exp () is index letter
Number operator;X (t) is matrix [x1, x2..., xh..., xm] t every trades vector, correspondence historical juncture t;X (j) is matrix
[x1, x2..., xh..., xm] jth every trade vector, it correspondence historical juncture j, be RBF center vector;||·
| | it is Euclidean Norm operator;δ press x (t) (t=1,2 ..., n) the Euclidean Norm maximum of all combination of two divided byMeter
Obtain, be the width of RBF;wj(j=1,2 ..., n) be regularization radial primary function network connection weight, it
By (w1, w2..., wn)T=(atj)+[y (1), y (2) ..., y (n)]TIt is calculated, wherein the general element of matrixThe plus sige broad sense inverse operator of superscript notation+be in algebra.
(6) according to the functional relationship of independent factor value to wind power standardized value, the wind power of basic point is predicted.First
By formulaCalculate the standardized value of basic point wind power, wherein x
(0)=[v '1(0), v '2(0) ..., v 'i..., v ' (0)5(0)][α1, a2..., ah..., am];Formula p (0)=y is pressed again
(0) σ (p)+ρ (p) is calculated, and obtains final product wind power prediction value p (0) of basic point.
Claims (3)
1. a kind of wind power forecasting method, it is characterised in that the method is comprised the following steps:
(1) to predict the time point of wind power as basic point, collection affects the historical data and wind of the meteorological factor of wind power
The matrix form of the historical data of electrical power, history of forming meteorological factor record and history wind power record;
(2) basic point meteorological factor record, history meteorological factor record and history wind power record are standardized respectively;
(3) standardization result with basic point meteorological factor record presses the standardization result that history meteorological factor is recorded as core
Euclidean distance filters out from small to large n record, obtains the matrix form of the meteorological neighbour's polymerization of basic point, and wherein n is positive integer,
N values are:5≤n≤10;
(4) to the meteorological factor record decorrelation dimensionality reduction in neighbour's polymerization, independent factor value matrix is obtained;
(5) according to the mapping relations of regularization radial primary function network independent factor value is set up to the letter of wind power standardized value
Number relation;
(6) prediction of basic point wind power is carried out by the functional relationship;
Gathering the historical data of meteorological factor and wind power in step (1) in 3~7 days before basic point, k values are 288~
672;Choose wind speed, wind direction, temperature, 5 meteorological factor of air pressure and humidity in step (1), wind speed, wind direction, temperature, air pressure and wet
Degree, is indicated successively using subscript numbering i=1,2 ..., 5;
Set up the matrix of basic point and history meteorological factor record:
Wherein:uiT () is record of the meteorological factor in moment t that numbering is i;T=0 and t=1,2 ..., k be respectively basic point and
History meteorological factor records the backward numbering at corresponding moment, and k is the number of history meteorological factor record;
Set up the column vector form of history wind power record:
P=[p (1), p (2) ... p (t) ..., p (k)]T
Wherein:P (t) is record of the wind power in historical juncture t;
Following two formulas standardization are pressed respectively to meteorological factor record and history wind power record in step (2):
v′i(t)=[ui(t)-ρ(ui)]/σ(ui)
Y (t)=[p (t)-ρ (p)]/σ (p)
Wherein:v′iT () is meteorological factor record uiThe standardization result of (t), p (ui) and σ (ui) it is respectively column vector uiMiddle element
Statistics average and standard deviation;Y (t) is the standardization result of history wind power record p (t), and ρ (p) and σ (p) are respectively
The statistics average and standard deviation of element in column vector p;
Standardization result [the v ' recorded with basic point meteorological factor in step (3)1(0), v '2(0) ..., v 'i..., v ' (0)5(0)]
For core, to t=1,2 ..., the standardization result of k history meteorological factor record press Euclidean distanceN record is filtered out from small to large, neighbour's polymerization is obtained, wherein 5≤n≤10;It is poly- to neighbour
Respectively recording the corresponding moment in conjunction renumbers in reverse order, obtains the matrix form recorded in neighbour's polymerization:
Wherein:viT () is record of the meteorological factor in moment t that numbering is i in neighbour's polymerization;
Decorrelation dimensionality reduction as the following formula is recorded to the meteorological factor in neighbour's polymerization in step (4), the matrix of independent factor value is obtained
Form:[x1, x2..., xh..., xm]=[v1, v2..., vi..., v5][a1, a2..., ah...., am]
Wherein:Subscript h is the numbering of independent factor;Column vector xhBe numbering be h independent factor in value not in the same time;Arrange to
Amount ahIt is matrix [v1, v2..., vi..., v5] the big characteristic roots of row correlation matrix h corresponding unit character vector, m
It is that to account for the ratio of characteristic root summation to the big characteristic root sums of m true more than 0.85 for the first big characteristic root of the row correlation matrix
Fixed positive integer.
2. wind power forecasting method according to claim 1, it is characterised in that functional relationship is described in step (5):
Wherein:Y (t) be historical juncture t (t=1,2 ..., n) the standardized value of wind power;X (t) and x (j) are respectively matrixes
[x1, x2..., xh..., xm] t rows, jth every trade vector, correspondence the historical juncture t and j;δ press x (t) (t=1,2 ..., n)
The Euclidean Norm maximum of all combination of two divided byIt is calculated;
wj(j=1,2 ..., n) press (w1, w2..., wn)T=(atjY)+[(1), y (2) ..., y (n)]TIt is calculated, wherein
3. wind power forecasting method according to claim 2, it is characterised in that press formula in step (6) firstThe standardized value of basic point wind power is calculated,
Wherein:X (0)=[v '1(0), v '2(0) ..., v 'i..., v ' (0)5(0)][a1, α2..., αh..., αm];Formula p is pressed again
(0)=y (0) σ (p)+ρ (p) obtains wind power prediction value p (0) of basic point.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410499648.8A CN104200290B (en) | 2014-09-26 | 2014-09-26 | Wind power forecast method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410499648.8A CN104200290B (en) | 2014-09-26 | 2014-09-26 | Wind power forecast method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104200290A CN104200290A (en) | 2014-12-10 |
CN104200290B true CN104200290B (en) | 2017-05-10 |
Family
ID=52085579
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410499648.8A Active CN104200290B (en) | 2014-09-26 | 2014-09-26 | Wind power forecast method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104200290B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105279582B (en) * | 2015-11-20 | 2019-01-04 | 水电十四局大理聚能投资有限公司 | Super short-period wind power prediction technique based on dynamic correlation feature |
CN109413149B (en) * | 2018-09-19 | 2021-05-18 | 上海哔哩哔哩科技有限公司 | Information distribution control method, system, server, and computer-readable storage medium |
CN112613674B (en) * | 2020-12-29 | 2024-03-08 | 国能日新科技股份有限公司 | Medium-and-long-term wind power generation capacity prediction method and device, electronic equipment and storage medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102521671A (en) * | 2011-11-29 | 2012-06-27 | 华北电力大学 | Ultrashort-term wind power prediction method |
CN104036328A (en) * | 2013-03-04 | 2014-09-10 | 横河电机株式会社 | Self-adaptive wind power prediction system and prediction method |
-
2014
- 2014-09-26 CN CN201410499648.8A patent/CN104200290B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102521671A (en) * | 2011-11-29 | 2012-06-27 | 华北电力大学 | Ultrashort-term wind power prediction method |
CN104036328A (en) * | 2013-03-04 | 2014-09-10 | 横河电机株式会社 | Self-adaptive wind power prediction system and prediction method |
Non-Patent Citations (1)
Title |
---|
基于天气分型的风电功率预测方法;王勃等;《电网技术》;20140131;第38卷(第1期);第93-98页 * |
Also Published As
Publication number | Publication date |
---|---|
CN104200290A (en) | 2014-12-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103065202B (en) | Wind power plant ultrashort term wind speed prediction method based on combination kernel function | |
CN102236795B (en) | Method for forecasting wind speed in wind power station | |
CN111754026B (en) | Photovoltaic power station group power prediction method and device, computer equipment and storage medium | |
CN104200290B (en) | Wind power forecast method | |
CN106505557A (en) | A kind of remote measurement misidentification method and device | |
CN101414751A (en) | Wind power forecasting system and method thereof, network system | |
CN107516145A (en) | A kind of multichannel photovoltaic power generation output forecasting method based on weighted euclidean distance pattern classification | |
CN101727538A (en) | Wind turbines input wind speed equivalent method taking into account influence of wind direction | |
CN103103570B (en) | Based on the aluminium cell condition diagnostic method of pivot similarity measure | |
CN107609707B (en) | Flood forecasting and disaster prevention decision method and system | |
CN104331572A (en) | Wind power plant reliability modeling method considering correlation between air speed and fault of wind turbine generator | |
CN107194600A (en) | A kind of electric load Seasonal Characteristics sorting technique | |
CN103514491A (en) | Power load forecasting method | |
CN104978459A (en) | IGBT remaining useful life prediction method | |
CN107403015A (en) | A kind of short-term luminous power Forecasting Methodology based on Time Series Similarity | |
CN108667069A (en) | A kind of short-term wind power forecast method returned based on Partial Least Squares | |
CN103996079A (en) | Wind power weighting predication method based on conditional probability | |
Yang et al. | Will the arid and semi-arid regions of Northwest China become warmer and wetter based on CMIP6 models? | |
CN116128141A (en) | Storm surge prediction method and device, storage medium and electronic equipment | |
CN114819374A (en) | Regional new energy ultra-short term power prediction method and system | |
CN102886384B (en) | Flatness defect identification method of 20-roller Sendzimir rolling mill based on support vector machine | |
Yang et al. | Short-term wind power prediction based on multi-parameters similarity wind process matching and weighed-voting-based deep learning model selection | |
CN106372440A (en) | Method and device for estimating self-adaptive robust state of distribution network through parallel computation | |
CN104850612A (en) | Enhanced cohesion hierarchical clustering-based distribution network user load feature classifying method | |
CN104834981B (en) | The main cause of short-term power station photovoltaic power implies type prediction technique |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |