CN106251242A - A kind of wind power output interval combinations Forecasting Methodology - Google Patents
A kind of wind power output interval combinations Forecasting Methodology Download PDFInfo
- Publication number
- CN106251242A CN106251242A CN201610642950.3A CN201610642950A CN106251242A CN 106251242 A CN106251242 A CN 106251242A CN 201610642950 A CN201610642950 A CN 201610642950A CN 106251242 A CN106251242 A CN 106251242A
- Authority
- CN
- China
- Prior art keywords
- wind
- output
- power output
- interval
- prediction
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 79
- 230000005611 electricity Effects 0.000 claims abstract description 22
- 230000005684 electric field Effects 0.000 claims description 54
- 238000005457 optimization Methods 0.000 claims description 10
- 230000007935 neutral effect Effects 0.000 claims description 9
- 238000013528 artificial neural network Methods 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 8
- 238000005259 measurement Methods 0.000 claims description 3
- 238000013459 approach Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 238000000611 regression analysis Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000000714 time series forecasting Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000005183 dynamical system Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003585 interneuronal effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 230000028161 membrane depolarization Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000007787 solid Substances 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
- 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)
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Water Supply & Treatment (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Wind Motors (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of wind power output interval combinations Forecasting Methodology, affected by natural wind fluctuations in wind speed for wind-power electricity generation and be there is the problem that the precision of prediction that bigger probabilistic problem and predicted time caused in advance declines, wind power output is carried out horizon prediction, and the combination realizing wind power output estimation range based on three kinds of methods adjusts, it is based respectively on wind speed changing ratio, based on predictive value rate of change, the forecast interval scope of wind power output is obtained based on three kinds of distinct methods of actual power optimal value, wind power output forecast interval optimum in selecting day part according to history wind power output data subsequently.The different parameters of the wind power output such as the inventive method combined wind velocity changing ratio, predictive value rate of change, actual power, and therefrom choose the forecast interval of optimum, it is advantageously implemented wind power output prediction more accurately.
Description
Technical field
The present invention relates to wind power generation output prediction field, in particular to based on the prediction of wind speed indeterminacy section
Wind power output is predicted, and sets up wind power output interval combinations forecast model based on three kinds of distinct methods.
Background technology
Wind-power electricity generation is the main path of clean electric power generation, but wind-power electricity generation is affected by natural wind wind speed, exists very
Big uncertainty.Wind-powered electricity generation prediction accurately is the effective way improving wind energy utilization efficiency, is also to ensure that under wind power integration electricity
The important means of net balance.Wind-power electricity generation enterprise can utilize the reasonable arrangement maintenance that predicts the outcome, and improves wind energy turbine set capacity coefficient,
Reduce cost of electricity-generating.Wind power output size is affected very big by the factor such as weather, season, and major influence factors has: wind speed;Empty
Air tightness;Wind direction.Wherein, on wind power output impact maximum is wind speed, and natural wind wind speed is inconstant, change in the short time
May be very big, it is the most disadvantageous for wind power output prediction.
Wind-powered electricity generation Forecasting Methodology currently mainly has: Time Series Forecasting Methods, regression analysis, Grey System Method,
Expert system approach, support vector machine etc..Time Series Forecasting Methods is according to historical data founding mathematical models, to following target
Data are predicted;Regression analysis, utilizes mathematical statistics to be analyzed the observation data of variable and determines and close between variable
System;Grey System Method is mainly used in the data modeling crossing sound pollution;The expert that expert system approach is stored by utilization knows
Know and experience comes reasoning decision-making;Support vector machine is based on statistical learning and structural risk minimization principle.In addition to the above methods,
Artificial neural network is also widely used in wind-powered electricity generation is predicted.Artificial neural network is to have self study, self-organizing and adaptive should be able to
One Kind of Nonlinear Dynamical System of power.The method utilizes past data to carry out the interneuronal weights of neutral net constantly
Revise, thus obtain ideal forecast model.The network system that a large amount of neurons are constituted can realize nonlinear mapping, divide
The functions such as class identification, optimization calculating, are suitable to solve wind power output forecasting problem.
Summary of the invention
Goal of the invention: in order to predict wind power output more accurately, the present invention proposes a kind of wind power output district
Between combined method, it is possible to increase the precision of wind power output interval prediction.
Technical scheme: for achieving the above object, the wind power output interval combinations method of the present invention, comprise the following steps:
(1) with the default cycle, the prediction period that wind power output is interval is divided for a certain area;
(2) wind power output historical data and the air speed data of wind energy turbine set each wind speed collection point of day part are gathered;
(3) for a certain prediction period, the air speed data collected is utilized to be based respectively on wind speed rate of change method, prediction
Wind power output estimation range is adjusted by value rate of change method, utilizes the wind power output historical data collected based on reality
Wind power output estimation range is optimized by power method;
(4) utilize the real data of the output of wind electric field of a period on this prediction period and prediction data to three kinds of methods
Precision of prediction is analyzed and selects the optimum prediction method of this prediction period, using corresponding wind power output forecast interval as group
Close the wind power output forecast interval of prediction.
Specifically, step (3) utilize the air speed data collected based on wind speed rate of change pair at a certain prediction period
Wind power output estimation range is adjusted, and comprises the following steps:
1) in the wind series of the upper period calculating this prediction period the air speed data of each moment and its previous moment it
Between wind speed changing ratio, formed the wind speed rate of change sequence of a upper period;
2) based on BP neural network computing, wind series of a period each moment in predicting this prediction period is utilized on this
Air speed data, obtains the wind speed rate of change sequence of this prediction period;
3) the wind speed rate of change in the wind speed rate of change sequence of this prediction period is divided wind speed rate of change according to equal length
Interval, adds up the wind speed rate of change interval obtaining maximum probability, by this interval to the wind speed rate of change fallen in each interval
Intermediate value as wind speed indeterminacy section coefficient, for this prediction period sometime, the actual measurement wind speed in this moment is taken advantage of
Carrying out wind speed with this indeterminacy section coefficient to adjust, the forecast interval obtaining subsequent time output of wind electric field is:
Ψ(v×(1-d),t)≤Pt+1≤Ψ(v×(1+d),t)
Ψ represents the nonlinear mapping relation between t wind speed and t+1 moment output of wind electric field, and this nonlinear mapping is closed
System is drawn by BP neural network computing.
Specifically, step (3) utilize the air speed data collected based on predictive value rate of change at a certain prediction period
Wind power output estimation range is adjusted, comprises the following steps:
1) rate of change of each moment output in the upper period calculating this prediction period;
2) the output rate of change and the output that utilized each moment in the upper period are predicted in this prediction period each
The output in moment;
3) calculate the output rate of change in each moment in this prediction period and divide output change according to equal length
Rate is interval, carries out the output rate of change fallen in each interval adding up the interval obtaining maximum probability, upper by this interval
Boundary is as power setting parameter δ;
4) utilize power setting parameter δ that output of wind electric field is adjusted, obtain a certain moment k+1 in this prediction period
The forecast interval Θ of output of wind electric field is:
Θ=[(1-δ) p (k+1), (1+ δ) p (k+1)].
Specifically, wind power output estimation range is optimized based on actual power in (3) by step at a certain prediction period,
Comprise the following steps:
1) utilize wind power output historical data and air speed data that BP neutral net is trained;
2) the BP neutral net trained is utilized to obtain the output prediction in each unknown moment in this prediction period successively
Value;
3) for a certain unknown moment of this prediction period, a upper moment output actual value in this unknown moment is calculated
And the difference between output predictive value, as parameters optimization δ in this moment, utilize the output to this moment of parameters optimization δ
Power prediction value is optimized, and obtains the forecast interval of the output of wind electric field in this moment, for the wind energy turbine set of certain unknown moment k
The forecast interval exerted oneself is:
Θk=[(1-δ) pk, (1+ δ) pk]。
Above-mentioned steps (4) utilizes real data and the forecast interval pair of the output of wind electric field of a period on this prediction period
The precision of prediction of three kinds of methods is analyzed and selects the optimum prediction method of this prediction period, particularly as follows:
Statistically in the period, the real data of each moment output of wind electric field respectively falls in the total of three kinds of method forecast intervals
Number, using method maximum corresponding for sum as the optimum prediction method of this prediction period.
Beneficial effect: the wind power output interval combinations method in the present invention be first based respectively on wind speed changing ratio, based on
Predictive value rate of change, neural network model and series model is utilized to obtain based on three kinds of distinct methods of actual power optimal value
The forecast interval scope of wind power output, wind power output prediction optimum in selecting day part according to history wind power output data subsequently
Interval, and day part prediction is chosen based on historical data the Forecasting Methodology of optimum, eliminate independent use wherein certain prediction
There is the shortcoming of certain limitation in method, is conducive to improving the precision of wind power output interval prediction;Wind power output in the present invention
Interval combinations method is uncertain for wind power output, replaces tradition wind-powered electricity generation deterministic forecast with interval prediction of exerting oneself, suitableeer
Answer the occasionality feature of wind power output;Research can be analyzed for different geographical environments, wind energy turbine set installation situation, have relatively
Strong adaptability, improvement and popularization for wind power output interval prediction method have preferable gain effect.
Accompanying drawing explanation
Fig. 1 is the flow chart of the wind power output interval combinations Forecasting Methodology in the present invention;
Fig. 2 is wind power output estimation range Optimizing Flow figure based on actual power;
Fig. 3 is to realize 95%~105% wind speed interval in prediction period based on wind speed changing ratio method in embodiment 1
Output of wind electric field predicts the outcome;
Fig. 4 is based on predictive value rate of change method wind power output interval prediction result in prediction period in embodiment 1;
Fig. 5 is that embodiment 1 wind power output based on actual power estimation range optimizes setting method in prediction period
Wind power output estimation range optimum results.
Detailed description of the invention
Below in conjunction with case study on implementation, the inventive method is further described.
As it is shown in figure 1, the wind power output interval combinations Forecasting Methodology in the present invention comprises the following steps:
(1) dividing the wind power output interval prediction period, the wind power output historical data and the wind energy turbine set that gather day part are each
The air speed data of wind speed collection point.
Output of wind electric field is mainly affected by wind speed, and wind speed exists bigger uncertainty, but the while of for areal
, there is certain rule in the wind speed of section (season), so it would be desirable to the period carrying out wind power output interval prediction is divided,
Being based primarily upon identical season (month) and identical period (date), different regions season division there are differences, and Time segments division is adopted
It it was a cycle with 2-4 days.
Such as: for a certain area, using No. 1 of August to No. 4 as a period, using No. 5 to No. 9 as a period,
Within 4 days, to be a cycle.
(2) utilize the air speed data of the wind energy turbine set each wind speed collection point gathered, be respectively adopted three kinds of distinct methods to wind
Electric field interval of exerting oneself is predicted, it was predicted that time use a period after the prediction of previous period data:
A) wind power output estimation range based on wind speed changing ratio is adjusted
The accuracy of output of wind electric field prediction is closely bound up with the accuracy of forecasting wind speed, in order to output of wind electric field scope
It is bound, first wind speed change can be predicted, realize output of wind electric field district based on wind speed variation prediction interval range
Between predict.
Definition wind speed rate of change:
Wind series v to a certain period1[k]={ v1,v2,v3...vk, will its sequence subsequent time v1[k] '=
{v2,v3...vk+1Every and v1[k]={ v1,v2,v3...vkSequence first does difference, then calculated difference relative to a upper moment
The ratio of wind speed, draws the wind speed rate of change sequence of this period:
ew[k]={ ew(2),ew(3),ew(4),...ew(k+1)};
BP neutral net is trained by the wind series utilizing a upper period of a certain prediction period, it was predicted that during this prediction
The air speed data in each moment in Duan, obtains the wind speed rate of change sequence of this prediction period further;
Wind speed rate of change in the wind speed rate of change sequence of this prediction period is divided wind speed rate of change district according to equal length
Between, such as: rate of change interval is divided into 0%-5%, several equal lengths such as 5%-10%, 10%-15% are interval, to falling respectively
Wind speed rate of change in interval carries out adding up that to obtain the wind speed rate of change interval of maximum probability be 5%-10%, by this interval
Between be worth 7.5% as wind speed indeterminacy section coefficient d, for this prediction period sometime, by the actual measurement wind speed in this moment
Being multiplied by this indeterminacy section coefficient d to carry out wind speed and adjust, the forecast interval obtaining subsequent time output of wind electric field is:
Ψ(v×(1-d),t)≤Pt+1≤ Ψ (v × (1+d), t),
Wherein, Pt+1I.e. representing the output of wind electric field of t subsequent time, Ψ represents t wind speed and t+1 moment wind-powered electricity generation
Field exert oneself between nonlinear mapping relation, this nonlinear mapping relation can be drawn by BP neural network computing.
B) wind power output estimation range based on predictive value rate of change is adjusted
Owing to the output of wind energy turbine set has bigger randomness, the sudden change of wind speed often causes the big of output
Amplitude of variation, the point prediction precision that these amplitudes of variation are bigger also can decline therewith.Basis at known output predictive value
On, it is possible to use wind power output estimation range is adjusted by the rate of change of output.
Define the output rate of change of a certain moment k:
The output rate of change and the output that utilize each moment in the period on a certain prediction period predict that this is pre-
Survey the output in each moment in the period;
Calculate the output rate of change in each moment in this prediction period and divide output rate of change according to equal length
Interval, carries out the output rate of change fallen in each interval adding up the interval obtaining maximum probability, by the upper bound in this interval
As power setting parameter δ, such as: output is divided into 0%-5%, several are isometric for 5%-10%, 10%-15% etc.
Degree interval, if falling at interval 5%-10% maximum probability, then setting parameter δ=10%;
Utilize power setting parameter δ that output of wind electric field is adjusted, obtain a certain moment k+1's in this prediction period
(output) forecast interval Θ of output of wind electric field is:
Θ=[(1-δ) p (k+1), (1+ δ) p (k+1)].
C) wind power output estimation range based on actual power optimizes
Above two method is all only based on historical data and carries out the wind power output prediction of certain period following, is not directed to
In prediction period, the actual of wind-powered electricity generation is exerted oneself.During prediction, can be in conjunction with the actual wind power output historical data correction obtained
The wind-powered electricity generation forecast interval of future time node.
In the case of known to a certain prediction period last time real output, utilize known time section [0, k-1]
Real output sequence p [k-1]={ r1,r2,r3...rk-1To BP neutral net output power prediction interval carry out excellent
Change.As in figure 2 it is shown, specifically include following steps:
1) utilize wind power output historical data and air speed data be trained BP neutral net obtaining input quantity wind speed and
Relation between output output of wind electric field;
2) the BP neutral net trained is utilized to obtain the output of wind electric field in each unknown moment in this prediction period successively pre-
Measured value;
3) for a certain unknown moment of this prediction period, by a upper moment output of wind electric field actual value in this unknown moment
With the difference between output of wind electric field predictive value is as parameters optimization δ in this moment, utilize the parameters optimization δ wind-powered electricity generation to this moment
Field predictive value of exerting oneself is optimized, and obtains the forecast interval of the output of wind electric field in this moment, for the wind-powered electricity generation of certain unknown moment k
The forecast interval exerted oneself in field is:
Θk=[(1-δ) pk, (1+ δ) pk]。
Such as: for moment k, when known past n time point output of wind electric field actual value sequence p [k-n, k-1]=
{pk-n,pk-n+1...pk-1And predictive value sequence r [k-n, k-1]={ rk-n, rk-n+1...rk-1Time, the most a certain moment k-1 wind
The electric field difference between actual value and predictive value of exerting oneself is Δ pk-1=pk-1-rk-1。
(3) for different situations, the interval combinations prediction of output of wind electric field is proposed on the basis of three kinds of methods.In advance
In the case of surveying the actual output of wind electric field the unknown of period in each moment, it is possible to use wind speed constant interval is predicted or utilizes predictive value to become
Rate setting range;In the case of known to the actual output of wind electric field of prediction period last time, available actual power is come excellent
Change the output prediction in follow-up unknown moment, and the output of wind electric field scope after being adjusted.
At certain prediction period, the wind power output of certain wind energy turbine set is predicted, is first respectively adopted above-mentioned three kinds of methods and carries out
Wind power output interval prediction, the real data of the output of wind electric field of a upper period and the prediction data precision of prediction to three kinds of methods
It is analyzed and selects the optimum prediction method of this prediction period, using corresponding wind power output forecast interval as combined prediction
Wind power output forecast interval.
Embodiment 1:
The present embodiment uses Matlab programming simulation platform, chooses certain wind energy turbine set period real data and carry out point
Analysis, is respectively adopted three kinds of methods based on actual, historical data and is predicted, and calculates three kinds of methods and imitates in the actual prediction of day part
Really, this period optimal forecast interval optimum interval Forecasting Methodology as this period is chosen.Day part forecast interval is with this type of
Push away, form overall interval prediction.
Choosing the historical data period is analyzed, and is respectively adopted three kinds of methods and is predicted, and to predicting the outcome with real
Border data compare, and choose this period optimal Forecasting Methodology (interval).
(1) wind power output estimation range setting method based on wind speed changing ratio is utilized, first to wind speed when each
Between the rate of change of point be predicted statistics, predicted that by BP neuroid the wind speed of existing wind speed subsequent time changes, and right
Amplitude of variation calculates ratio on the basis of existing wind speed, adds up all ratios, counts the change of subsequent time wind speed
Maximum possible is interval.
It is that previous moment wind speed ratio major part (probability is 48%) is positioned at 4%~6% through statistics subsequent time wind speed
Wind speed rate of change is interval, then wind speed indeterminacy section coefficient d takes 5%.Air speed data thus can be given upper lower limit value, thus obtain
The scope exerted oneself to prediction.By this result to artificial neural network input measuring wind speed value 95%~105% as wind
Speed input bound, the wind power output area Results obtained is as shown in Figure 3.
Solid black lines is actual output of wind electric field, and gray area is wind power output wind speed bound obtained as input
Upper lower limit value, it is seen that in the range of actual wind power output value is predominantly located in light areas, showing can when wind speed fluctuates widely
Use the mode predicting upper lower limit value of exerting oneself.The forecast interval of the wind power output that model obtains, is fluctuating violent position still
So there is certain error, but be suitable for the uncertainty of wind power output than the method for prediction wind power output of the prior art,
Uncertainty in the wind power output bound interval balance wind-powered electricity generation prediction thus used, chooses offer for system reserve capacity
Fully foundation.
(2) wind-powered electricity generation output of wind electric field is pre-to use wind power output estimation range setting method based on predictive value rate of change to consider
The rate of change of measured value, by selecting rational mapping relations to reflect the uncertainty of wind power output: when output of wind electric field changes
When rate is in maximum probability in the range of 0-5%, corresponding output of wind electric field scope is the 95%~105% of predictive value;Work as wind energy turbine set
When rate of change of exerting oneself is in maximum probability in the range of 5-10%, corresponding output of wind electric field scope is predictive value 90%~
110%;By that analogy, in the present embodiment in the range of output of wind electric field rate of change is in 15-20% maximum probability, maximum wind
Field range set of exerting oneself is 80%~120%, and obtain exerts oneself area Results as shown in Figure 4.
Result from Fig. 4 is it can be seen that change relatively large point for wind power output, and output of wind electric field actual value is permissible
Preferably fall in setting range;Interval effectively reducing of adjusting can also be realized going out the less point of fluctuation.But still deposit
Actual power at part-time point exceeds the situation of setting range.
(3) wind power output estimation range optimization method based on actual power is used to realize output of wind electric field scope whole
Fixed, when the actual output of wind electric field that situation is known wind energy turbine set last time in predicted time section, an above moment wind-powered electricity generation
The difference exerted oneself between actual value and predictive value in field, as input, obtains the output of wind electric field range intervals of correspondence.Difference range
Corresponding with the relation of interval parameter of adjusting as shown in table 1 below:
Table 1 difference range and interval parameter table of adjusting
Result of adjusting from the interval of wind power output estimation range optimization method based on actual power shown in Fig. 5 can be sent out
Existing, in the case of the actual output of wind electric field of known wind energy turbine set, preferably output area can be made and adjusting, obtain preferably
Output of wind electric field forecast interval.Depolarization minority wind-powered electricity generation actual output of wind electric field value beyond estimation range, most wind energy turbine set
Actual value of exerting oneself all fall within the scope of.
Table 2 each method wind power output forecast interval accuracy rate is added up
Comparative result by the comprehensive three of table 2: use the wind power output estimation range side of adjusting based on wind speed changing ratio
Method 5% is floated and is inputted, and the data point beyond setting range has 54 (144, strong points of sum);Use based on predictive value rate of change
Wind power output estimation range setting method have 70 beyond the data point of setting range;Wind-powered electricity generation based on actual power is used to go out
Power estimation range optimization method has 32 beyond the data point of setting range.Result shows when the unknown real-time output of wind electric field of wind-powered electricity generation
In the case of can use Forecasting Methodology based on forecasting wind speed constant interval, and in the situation of the actual output of wind electric field of known wind-powered electricity generation
Under, use instantaneous value forecast interval relatively predictive value rate of change setting method of adjusting can obtain and preferably predict the outcome.Combine
Output of wind electric field scope is adjusted by three kinds of methods, be based primarily upon wind speed changing ratio prediction adjust, predictive value rate of change whole
Determine scope and instantaneous value correction setting range.Above-mentioned three kinds of methods are used to predict respectively somewhere wind power output sectional,
And investigate each method accuracy at different periods based on history real data, the Forecasting Methodology choosing day part optimum carries out pre-
Survey, and form final prediction and exert oneself interval.
Claims (5)
1. a wind power output interval combinations Forecasting Methodology, it is characterised in that comprise the following steps:
(1) with the default cycle, the prediction period that wind power output is interval is divided for a certain area;
(2) wind power output historical data and the air speed data of wind energy turbine set each wind speed collection point of day part are gathered;
(3) for a certain prediction period, the air speed data collected is utilized to be based respectively on wind speed rate of change method, predictive value change
Wind power output estimation range is adjusted by rate method, utilizes the wind power output historical data collected based on actual power
Wind power output estimation range is optimized by method;
(4) real data and the prediction data prediction to three kinds of methods of the output of wind electric field of a period on this prediction period are utilized
Precision is analyzed and selects the optimum prediction method of this prediction period, and corresponding wind power output forecast interval is pre-as combination
The wind power output forecast interval surveyed.
Wind power output interval combinations Forecasting Methodology the most according to claim 1, it is characterised in that a certain in step (3)
Prediction period utilizes the air speed data that collected to adjust wind power output estimation range based on wind speed rate of change, including with
Lower step:
1) in the wind series of the upper period calculating this prediction period between each moment and the air speed data of its previous moment
Wind speed changing ratio, formed the wind speed rate of change sequence of a upper period;
2) based on BP neural network computing, wind series of period wind speed in each moment in predicting this prediction period is utilized on this
Data, obtain the wind speed rate of change sequence of this prediction period;
3) the wind speed rate of change in the wind speed rate of change sequence of this prediction period is divided wind speed rate of change according to equal length interval,
The wind speed rate of change fallen in each interval is added up the wind speed rate of change interval obtaining maximum probability, by the centre in this interval
The actual measurement wind speed in this moment, as wind speed indeterminacy section coefficient, for this prediction period sometime, is multiplied by this not by value
Determining that Interval carries out wind speed and adjusts, the forecast interval obtaining subsequent time output of wind electric field is:
Ψ(v×(1-d),t)≤Pt+1≤Ψ(v×(1+d),t)
Ψ represents the nonlinear mapping relation between t wind speed and t+1 moment output of wind electric field, and this nonlinear mapping relation is led to
Cross BP neural network computing to draw.
Wind power output interval combinations Forecasting Methodology the most according to claim 1, it is characterised in that a certain in step (3)
Prediction period utilizes the air speed data collected to adjust wind power output estimation range based on predictive value rate of change, including
Following steps:
1) rate of change of each moment output in the upper period calculating this prediction period;
2) the output rate of change in each moment in the upper period and output was utilized to predict each moment in this prediction period
Output;
3) calculate the output rate of change in each moment in this prediction period and divide output rate of change district according to equal length
Between, carry out the output rate of change fallen in each interval adding up the interval obtaining maximum probability, the upper bound in this interval is made
For power setting parameter δ;
4) utilize power setting parameter δ that output of wind electric field is adjusted, obtain a certain moment k+1 wind-powered electricity generation in this prediction period
The forecast interval Θ exerted oneself in field is:
Θ=[(1-δ) p (k+1), (1+ δ) p (k+1)].
Wind power output interval combinations Forecasting Methodology the most according to claim 1, it is characterised in that a certain in step (3)
Wind power output estimation range is optimized by prediction period based on actual power, comprises the following steps:
1) utilize wind power output historical data and air speed data that BP neutral net is trained;
2) the BP neutral net trained is utilized to obtain the output predictive value in each unknown moment in this prediction period successively;
3) for a certain unknown moment of this prediction period, the upper moment output actual value in this unknown moment and defeated is calculated
Go out the difference between power prediction value, as parameters optimization δ in this moment, utilize the parameters optimization δ output to this moment
Predictive value is optimized, and obtains the forecast interval of the output of wind electric field in this moment, for the output of wind electric field of certain unknown moment k
Forecast interval be:
Θk=[(1-δ) pk, (1+ δ) pk]。
Wind power output interval combinations Forecasting Methodology the most according to claim 1, it is characterised in that utilizing in step (4) should
On prediction period, the precision of prediction of three kinds of methods is analyzed by real data and the forecast interval of the output of wind electric field of a period
And select the optimum prediction method of this prediction period, particularly as follows:
Statistically in the period, the real data of each moment output of wind electric field respectively falls in the sum of three kinds of method forecast intervals, will
The maximum corresponding method of sum is as the optimum prediction method of this prediction period.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610642950.3A CN106251242B (en) | 2016-08-08 | 2016-08-08 | Wind power output interval combination prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610642950.3A CN106251242B (en) | 2016-08-08 | 2016-08-08 | Wind power output interval combination prediction method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106251242A true CN106251242A (en) | 2016-12-21 |
CN106251242B CN106251242B (en) | 2019-12-10 |
Family
ID=58079181
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610642950.3A Active CN106251242B (en) | 2016-08-08 | 2016-08-08 | Wind power output interval combination prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106251242B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106875033A (en) * | 2016-12-26 | 2017-06-20 | 华中科技大学 | A kind of wind-powered electricity generation cluster power forecasting method based on dynamic self-adapting |
CN108446506A (en) * | 2018-03-26 | 2018-08-24 | 东北大学 | A kind of uncertain system modeling method based on section Feedback Neural Network |
CN108876060A (en) * | 2018-08-01 | 2018-11-23 | 国网吉林省电力有限公司长春供电公司 | A kind of sample collection wind power output probability forecasting method based on big data |
CN109449929A (en) * | 2018-11-22 | 2019-03-08 | 南方电网科学研究院有限责任公司 | Distributed generation resource influences prediction and evaluation method and product to distribution network feeder utilization rate |
CN109960818A (en) * | 2017-12-22 | 2019-07-02 | 北京金风慧能技术有限公司 | Generate the method and device of the simulation air speed data of wind power plant |
CN112036607A (en) * | 2020-07-30 | 2020-12-04 | 南方电网科学研究院有限责任公司 | Wind power output fluctuation prediction method and device based on output level and storage medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110078226A1 (en) * | 2009-09-30 | 2011-03-31 | International Business Machines Corporation | Sparse Matrix-Vector Multiplication on Graphics Processor Units |
CN105552970A (en) * | 2016-02-25 | 2016-05-04 | 华北电力科学研究院有限责任公司 | Method and apparatus for improving accuracy of predicting power of wind power station |
-
2016
- 2016-08-08 CN CN201610642950.3A patent/CN106251242B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110078226A1 (en) * | 2009-09-30 | 2011-03-31 | International Business Machines Corporation | Sparse Matrix-Vector Multiplication on Graphics Processor Units |
CN105552970A (en) * | 2016-02-25 | 2016-05-04 | 华北电力科学研究院有限责任公司 | Method and apparatus for improving accuracy of predicting power of wind power station |
Non-Patent Citations (1)
Title |
---|
李艳晴等: "基于卡尔曼滤波的迟滞神经网络风速序列预测", 《北京科技大学学报》 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106875033A (en) * | 2016-12-26 | 2017-06-20 | 华中科技大学 | A kind of wind-powered electricity generation cluster power forecasting method based on dynamic self-adapting |
CN106875033B (en) * | 2016-12-26 | 2020-06-02 | 华中科技大学 | Wind power cluster power prediction method based on dynamic self-adaption |
CN109960818A (en) * | 2017-12-22 | 2019-07-02 | 北京金风慧能技术有限公司 | Generate the method and device of the simulation air speed data of wind power plant |
CN109960818B (en) * | 2017-12-22 | 2023-06-20 | 北京金风慧能技术有限公司 | Method and device for generating simulated wind speed data of wind power plant |
CN108446506A (en) * | 2018-03-26 | 2018-08-24 | 东北大学 | A kind of uncertain system modeling method based on section Feedback Neural Network |
CN108446506B (en) * | 2018-03-26 | 2021-06-08 | 东北大学 | Uncertain system modeling method based on interval feedback neural network |
CN108876060A (en) * | 2018-08-01 | 2018-11-23 | 国网吉林省电力有限公司长春供电公司 | A kind of sample collection wind power output probability forecasting method based on big data |
CN108876060B (en) * | 2018-08-01 | 2021-05-11 | 国网吉林省电力有限公司长春供电公司 | Big data based prediction method for wind power output probability of sample collection |
CN109449929A (en) * | 2018-11-22 | 2019-03-08 | 南方电网科学研究院有限责任公司 | Distributed generation resource influences prediction and evaluation method and product to distribution network feeder utilization rate |
CN112036607A (en) * | 2020-07-30 | 2020-12-04 | 南方电网科学研究院有限责任公司 | Wind power output fluctuation prediction method and device based on output level and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN106251242B (en) | 2019-12-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106251242A (en) | A kind of wind power output interval combinations Forecasting Methodology | |
Wang et al. | Research and application of a combined model based on multi-objective optimization for multi-step ahead wind speed forecasting | |
Liu et al. | Forecasting power output of photovoltaic system using a BP network method | |
De Giorgi et al. | Error analysis of short term wind power prediction models | |
Gomes et al. | Wind speed and wind power forecasting using statistical models: autoregressive moving average (ARMA) and artificial neural networks (ANN) | |
CN103268366B (en) | A kind of combination wind power forecasting method suitable for distributing wind power plant | |
CN109657881A (en) | A kind of neural network photovoltaic power generation prediction technique and system suitable for small sample | |
Zhang et al. | GEFCom2014 probabilistic solar power forecasting based on k-nearest neighbor and kernel density estimator | |
CN105512745A (en) | Wind power section prediction method based on particle swarm-BP neural network | |
CN102478584B (en) | Wind power station wind speed prediction method based on wavelet analysis and system thereof | |
CN105046953A (en) | Short-time traffic-flow combination prediction method | |
CN101893674A (en) | Pollution flashover index forecasting method for regional power grid | |
CN104408562A (en) | Photovoltaic system generating efficiency comprehensive evaluation method based on BP (back propagation) neural network | |
Kolhe et al. | GA-ANN for short-term wind energy prediction | |
CN104732300A (en) | Neural network wind power short-term forecasting method based on fuzzy partition theory | |
CN104598765A (en) | Building energy consumption prediction method based on elastic adaptive neural network | |
Tarade et al. | A comparative analysis for wind speed prediction | |
CN104820877A (en) | Photovoltaic system generation power prediction method based on cloud adaptive PSO-SNN | |
CN103793887A (en) | Short-term electrical load on-line predicting method based on self-adaptation enhancing algorithm | |
Liao et al. | Ultra-short-term interval prediction of wind power based on graph neural network and improved bootstrap technique | |
CN107832881A (en) | Consider the wind power prediction error assessment method of load level and wind speed segmentation | |
Yang et al. | Sub-minute probabilistic solar forecasting for real-time stochastic simulations | |
CN105956708A (en) | Grey correlation time sequence based short-term wind speed forecasting method | |
CN103745274A (en) | Short-term power prediction method applied to dispersed wind power | |
CN113610285A (en) | Power prediction method for distributed wind power |
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 |