CN103577890B - Based on the matched cluster point wind power forecasting method of tuple - Google Patents
Based on the matched cluster point wind power forecasting method of tuple Download PDFInfo
- Publication number
- CN103577890B CN103577890B CN201310451220.1A CN201310451220A CN103577890B CN 103577890 B CN103577890 B CN 103577890B CN 201310451220 A CN201310451220 A CN 201310451220A CN 103577890 B CN103577890 B CN 103577890B
- Authority
- CN
- China
- Prior art keywords
- tuple
- sequence
- time
- wind power
- data
- 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.)
- Expired - Fee Related
Links
- 238000013277 forecasting method Methods 0.000 title abstract 3
- 238000000034 method Methods 0.000 claims abstract description 51
- 230000009466 transformation Effects 0.000 claims abstract description 8
- 238000012545 processing Methods 0.000 claims abstract description 5
- 230000006835 compression Effects 0.000 claims description 23
- 238000007906 compression Methods 0.000 claims description 23
- 238000004364 calculation method Methods 0.000 claims description 12
- 230000001186 cumulative effect Effects 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 7
- 230000003139 buffering effect Effects 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 5
- 238000005259 measurement Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 abstract description 2
- 230000005611 electricity Effects 0.000 abstract 4
- 238000011161 development Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 238000013179 statistical model Methods 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 241001123248 Arma Species 0.000 description 1
- 241000764238 Isis Species 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002688 persistence Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
- Y02E40/70—Smart grids as climate change mitigation technology in the energy generation sector
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses one kind to be based on the matched cluster point wind power forecasting method of tuple, comprising the following steps: the active power data of input wind-powered electricity generation cluster point；Piece-wise linearization processing is carried out using wind-powered electricity generation data of the piecewise-linear techniques to wind-powered electricity generation cluster point, to treated as a result, being indicated using tuple sequence；Determine tuple sequence S to be matched0And extract the tuple sequence composition tuple sequence collection S of formationset, tuple sequence S to be matched0It is with current time T1It is pushed forward complete day d, time interval [T1‑d,T1] in wind power output data tuple sequence；Tuple sequence collection S is found out using tuple dynamic distortion matching algorithmsetNeutralize tuple sequence S to be matched0Target tuple sequence S the most matched1；Calculate tuple sequence S to be matched0With target tuple sequence S1Between transformation coefficient；Using obtained transformation coefficient to target tuple sequence S1Historical data later carries out operation；Export cluster point wind power prediction value.Solve the problems, such as that wind-powered electricity generation cluster point power can not be effectively predicted in existing wind power forecasting method.
Description
Technical Field
The invention relates to the field of wind power generation, in particular to a cluster point wind power prediction method based on tuple matching.
Background
With the vigorous development of new energy in the world and the development of scientific technology in a new and different day by day, the wind power industry in China develops rapidly, and the development of the wind power prediction technology is closely related to the development of national economy. The method has very important significance in researching the short-term change condition of the wind power cluster point. In the power market mode, especially after large-scale wind power is accessed, the operation mode and mode of a power grid are changed greatly, a power grid dispatcher needs to adjust the operation mode of the power grid in time according to market requirements and the power generation characteristics of the wind power, and the power grid dispatching method is particularly important for simply, accurately and quickly predicting the wind power of a power grid cluster point when economic load distribution and real-time dispatching are carried out.
The traditional wind power prediction method in the power system mainly comprises two types: the method is mainly applied to power prediction of a wind power plant layer. The class can be classified into a statistical model and a physical model. The statistical model is a mapping relation between input and output established by using a statistical tool, and the specific method comprises an autoregressive technology, a support vector machine, a neural network and the like. The physical model substantially improves the resolution of the meteorological prediction model, so that the meteorological prediction model can accurately predict single-point weather; and the other method is a prediction method without using weather forecast, which is called a historical data-based prediction method, and the method predicts the wind power only according to historical data, and comprises a Kalman filtering method, a persistence algorithm, an ARMA algorithm and a linear regression model.
The two methods have defects in the aspect of wind power prediction of wind power cluster points. Because the wind power of the cluster point is the centralized embodiment of the power of the connected wind power plants, the meteorology of each wind power plant is different, and there is no basis for predicting the wind power output of the cluster point by using the forecast result of the meteorology of the cluster point, the method of the first class is difficult to apply in the aspect of forecasting the wind power of the cluster point. In addition, at present, research objects of the wind power prediction method based on historical data are always on a wind power plant layer, cluster points are rarely researched, and the method has some defects in algorithm efficiency.
Disclosure of Invention
The invention aims to provide a cluster point wind power prediction method based on tuple matching to solve the problem that the conventional wind power prediction method cannot effectively predict the power of wind power cluster points.
In order to achieve the purpose, the invention adopts the technical scheme that:
a cluster point wind power prediction method based on tuple matching comprises the following steps:
step 1: inputting active power data of all historical wind power outputs of the wind power cluster points;
step 2: carrying out piecewise linearization processing on wind power historical output data of the wind power cluster points by utilizing a piecewise linearization method, and expressing a processed result by adopting a tuple sequence;
and step 3: determining a sequence S of tuples to be matched0And extracting the tuple sequence formed in the step 2 to form a tuple sequence set SsetThe tuple sequence S to be matched0Is the current time T1Push forward a complete day d, time interval [ T1-d,T1]The tuple sequence of the internal wind power output data;
and 4, step 4: solving the tuple sequence set S by using a tuple dynamic distortion matching algorithmsetNeutralizing tuple sequence S to be matched0Best matched target tuple sequence S1；
And 5: calculating a tuple sequence S to be matched0And a sequence of target tuples S1The transform coefficients between;
step 6: utilizing the transformation coefficient obtained in the step 5 to pair the target tuple sequence S1Calculating the subsequent historical data;
and 7: and outputting the predicted value of the wind power of the cluster point.
According to the preferred embodiment of the invention, the piecewise linearization method is a hybrid piecewise linearization algorithm adopting a SWAB algorithm, the SWAB algorithm is a combination of a classic slipping-Window algorithm and a Bottom-Up algorithm, the buffering area is dynamically updated by the slipping-Window algorithm, and the Bottom-Up algorithm carries out piecewise linearization on data in the buffering area again.
According to a preferred embodiment of the present invention, the tuple sequence refers to a method for representing piecewise linearized data, the tuple refers to a set of discrete data, and a line segment is represented by using a 4-tuple L (x, y, k, Δ x), where x is an abscissa of any point on a straight line, y is an ordinate of the point, k is a slope of the straight line, and Δ x is a projection length of the line segment on an x-axis; the tuple sequence is an ordered set of tuples, and for the history data after m segments, the tuple sequence is expressed as:
S＝{L1(x1,y1,k1,Δx1),…,Lm(xm,ym,km,Δxm)}
according to the preferred embodiment of the present invention, the tuple sequence set SsetIn the historical wind power output data of cluster points, the time periods are respectively
[T1-2d,T1-d],[T1-3d,T1-2d],…,[T1-nd,T1-(n-1)d]…, the data are respectively processed by segment linearization and tuple representation to form tuple sequence set.
According to the preferred embodiment of the present invention, the tuple dynamic twist matching algorithm is as follows:
the tuple dynamic distortion matching algorithm is a nonlinear warping algorithm combining time warping, tuple representation and distance measurement calculation, a mapping function phi () is found, and a tuple is testedIs nonlinearly mapped to the reference tupleAnd such that the function satisfies:
wherein,are respectively a test tuple sequence SaAnd a reference tuple sequence SbThe number of n, m tuples in (a),is the distance between two tuples, D is the distance between two tuple sequences under the optimal time warping condition, N is the tuple sequence SaThe number of tuples in (1);
tuple metric function of an algorithmThe calculation is as follows:
is provided with a tuple sequence Sa，Sb，，And
are respectively Sa,SbN, m segmented tuples of (S)a,SbRespectively has a sequence length of delta A, delta B, w0For the weighting coefficients, the tuple distance calculation function is as follows
Wherein the weighting coefficient function:
magnitude distance function:
andexpressed as a sequence S of test tuplesaAnd a reference tuple sequence SbThe slope of (a) of (b) is,
the cumulative distance of the minimum path is derived using a cumulative distance calculation formula as follows:
d is the distance of the two tuple sequence under the optimal time warping condition.
According to a preferred embodiment of the invention, said coefficient of variation is the temporal compression ratio λtimeTime offset amount deltatimeAnd amplitude compression ratio lambdapowerThe method comprises the following steps:
obtaining a target tuple sequence S according to the matching path1And tuple sequence S to be matched0One-to-one correspondence of tuples inThe system has a total of n tuple pairsAnd a sequence of tuples of the respective
S1,S2I, j tuples of (1);
the time compression ratio lambdatime：
The time compression rate is characterized by the average compression degree between the two tuple sequences on a time axis;
offset in time Δtime
The time offset represents the dislocation degree of the two tuple sequences on the time axis;
amplitude compression ratio lambdapower
The amplitude compression rate characterizes the average degree of compression of the two tuple of sequences over the amplitude.
According to the preferred embodiment of the present invention, the historical data after the target tuple sequence refers to when the time interval of the target tuple sequence is [ T ]1,T2]Time-series and cluster point wind power output historical dataMiddle T2All data after the time.
According to a preferred embodiment of the present invention, the operation of the change coefficient on the history data is: the method is characterized in that after two adjacent data points of the historical data are connected, the function of the historical data is expressed as
H (t), t is more than or equal to 0, the operation result is:
H′(t)＝λpower2H(λtime·(t-Δtime)),t≥0。
according to the preferred embodiment of the present invention, the output cluster point wind power prediction value refers to:
discretizing the H ' (T) at the sampling time interval DeltaT of the original historical data, and taking data points H ' (0), H ' (1), …, H ' (T/DeltaT) with a required time length T from the discretized H ' (n) as prediction data.
The technical scheme of the invention has the following beneficial effects:
according to the technical scheme, a cluster point wind power prediction model based on historical data is established by adopting a tuple matching method, historical operation data are fully utilized by utilizing the idea that phenomena under the same mechanism are approximately reproduced in the prediction theory, future wind power output can be rapidly predicted, and the prediction precision can be continuously improved along with the continuous enrichment of a historical operation library, so that the problem that the power of a wind power cluster point cannot be effectively predicted by the conventional wind power prediction method is solved.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
FIG. 1 is a flowchart of a cluster point wind power prediction method based on tuple matching according to an embodiment of the present invention;
FIG. 2 is a flow chart of a piecewise linearization algorithm according to an embodiment of the invention;
FIG. 3 is a schematic diagram of a SWAB piecewise linearization algorithm according to an embodiment of the invention;
FIG. 4 is a diagram illustrating a matching path of a tuple dynamic warp matching algorithm according to an embodiment of the present invention;
FIG. 5 illustrates data to be matched and its piecewise linearization effect according to an embodiment of the present invention;
FIGS. 6a to 6d are diagrams illustrating the result of tuple dynamic warp matching according to the embodiment of the present invention;
fig. 7 is a comparison graph of the active power flow predicted value and the actual value according to the embodiment of the present invention.
Detailed Description
The preferred embodiments of the present invention will be described in conjunction with the accompanying drawings, and it will be understood that they are described herein for the purpose of illustration and explanation and not limitation.
As shown in fig. 1, a cluster point wind power prediction method based on tuple matching includes the following steps:
step 101: inputting active power data of all historical wind power outputs of the wind power cluster points;
step 102: carrying out piecewise linearization processing on wind power historical output data of the wind power cluster points by utilizing a piecewise linearization method, and expressing a processed result by adopting a tuple sequence;
step 103: determining a sequence S of tuples to be matched0And extracting the tuple sequence formed in step 102 to form a tuple sequence set SsetSequence of tuples to be matched S0Is the current time T1Push forward a complete day d, time interval [ T1-d,T1]The tuple sequence of the internal wind power output data;
step 104:solving tuple sequence set S by using tuple dynamic distortion matching algorithmsetNeutralizing tuple sequence S to be matched0Best matched target tuple sequence S1；
Step 105: calculating a tuple sequence S to be matched0And a sequence of target tuples S1The transform coefficients between;
step 106: using the transformation coefficient obtained in step 105 to pair the target tuple sequence S1Calculating the subsequent historical data;
step 107: and outputting the predicted value of the wind power of the cluster point.
As shown in fig. 2 and fig. 3, the piecewise linearization method is a hybrid piecewise linearization algorithm that adopts a swap algorithm, the swap algorithm is a combination of a classic slipping-Window algorithm and a Bottom-Up algorithm, the buffering area is dynamically updated by the slipping-Window algorithm, and the Bottom-Up algorithm performs piecewise linearization on data in the buffering area again.
As shown in fig. 4, a tuple sequence refers to a representation method of data linearized in segments, a tuple refers to a collection of discrete data, and a 4-tuple L (x, y, k, Δ x) is used to represent a line segment, where x is an abscissa of any point on a straight line, y is an ordinate of the point, k is a slope of the straight line, and Δ x is a projection length of the line segment on an x-axis; the tuple sequence is an ordered set of tuples, and for m-segmented historical data, the m-segmented historical data is represented by the tuple sequence, wherein 4-tuples comprise 4 tuples and less than 4 tuples, namely 3 tuples, 2 tuples and 1 tuples.
S＝{L1(x1,y1,k1,Δx1),…,Lm(xm,ym,km,Δxm)}
Tuple sequence set SsetIn the historical wind power output data of cluster points, the time period is T1-2d,T1-d],[T1-3d,T1-2d],…,[T1-nd,T1-(n-1)d]… are respectively represented by piecewise linearization and tupleA set of tuple sequences is composed.
The tuple dynamic twist matching algorithm is specifically derived as follows:
the tuple dynamic distortion matching algorithm is a nonlinear warping technique which combines time warping, tuple representation and distance measurement calculation. It finds a mapping function phi (phi), which is a non-linear mapping, and the specific implementation of the function is determined by the calculation of cumulative distance, and the test tuple sequence isIs nonlinearly mapped to the reference tupleAnd such that the function satisfies:
wherein,are respectively a test tuple sequence SaAnd a reference tuple sequence SbThe number of n, m tuples in (a),is the distance between two tuples, D is the distance of the two tuple sequence under the optimal time warping condition, N is SaThe number of tuples.
The tuple metric function of the algorithm refers to:
is provided with a tuple sequence Sa，Sb And is divided intoIs other than Sa,SbN, m segmented tuples of (S)a,SbIs long in sequenceThe degrees are Δ a, Δ B, respectively. The tuple distance calculation function is as follows
Wherein the weighting coefficient function:
magnitude distance function:
andexpressed as a sequence S of test tuplesaAnd a reference tuple sequence SbSlope of (1), w0Are weight coefficients.
The dynamic distortion matching process of the algorithm is the application of the classic dynamic time distortion matching algorithm in the tuple matching problem. Through the dislocation matching of the tuple sequence, the optimal matching sequence in a certain distortion range is found, and compared with the traditional matching algorithm based on the Euclidean distance, the method is more robust to the offset and amplitude expansion of the time sequence. The algorithm strives to find a matching optimal path to minimize the cumulative distance of the path, and the cumulative distance calculation formula is as follows:
the path constraints of the algorithm are as follows: in order to solve the problem that the traditional dynamic time warping algorithm requires that the first tuples must be matched, the search starts from a certain node in (1,1), (1,2), (1,3), (2,1) and (3, 1); the next lattice point of the current point (n, m) can only be one of (n, m), (n +1, m +1) and (n, m + 1); assuming the path width constraint is r, the two tuple sequence number difference | i-j | is required to satisfy
And | i-j | is less than or equal to r + | M-N | wherein M, N is the number of tuples of the two sequences respectively.
Target tuple sequence S1Refers to the tuple sequence set S obtained by the tuple matching algorithmsetMiddle and tuple sequence S to be matched0The most matched tuple sequence.
The coefficient of variation being the temporal compressibility λtimeTime offset amount deltatimeAnd amplitude compression ratio lambdapowerThe method comprises the following steps:
obtaining a target tuple sequence S according to the matching path1And tuple sequence S to be matched0One-to-one correspondence of tuples inThe system has a total of n tuple pairsAnd a sequence of tuples of the respective
S1,S2I, j tuples of (1);
the time compression ratio lambdatime：
The time compression rate is characterized by the average compression degree between the two tuple sequences on a time axis;
offset in time Δtime
The time offset represents the dislocation degree of the two tuple sequences on the time axis;
amplitude compression ratio lambdapower
The amplitude compression rate characterizes the average degree of compression of the two tuple of sequences over the amplitude.
The historical data after the target tuple sequence refers to the time interval when the target tuple sequence is [ T ]1,T2]T in time and cluster point wind power output historical data2All data after the time.
Operation of the coefficient of variation on the historical data: after every two adjacent data points of the historical data are connected, the function of the historical data is expressed as H (t), and if t is more than or equal to 0, the operation result is as follows:
H′(t)＝λpower·H(λtime·(t-Δtime)),t≥0。
outputting the predicted value of the wind power of the cluster point refers to:
discretizing the H ' (T) at the sampling time interval DeltaT of the original historical data, and taking data points H ' (0), H ' (1), …, H ' (T/DeltaT) with a required time length T from the discretized H ' (n) as prediction data.
The technical scheme of the invention selects 2011 annual Gansu power grid cluster Dunhuang change 2# main transformer 330 side active power flow to establish a database, and the side is only connected with a plurality of wind power plants without other forms of power supplies and loads.
The piecewise linearization algorithm of the invention is adopted to carry out piecewise linearization processing and is expressed by tuples to form a matching history database. Assuming that the current day is 2011, 9 and 24 days, determining the active power flow data of the 2011, 9 and 25 days by using a tuple matching algorithm, and performing comparative analysis by using actual data to verify the text method. The analysis was as follows:
as shown in fig. 5, 2011 shows the active power flow of 24 th day in 9 months, and the broken line in the figure shows the effect of the piecewise linearization representation. And matching the data with a historical database, and taking four most matched records, namely 7 days 7 and 7 months in 2011, 11 days 4 and 11 months in 2011, 2 days 9 and 2 months in 2011 and 12 days 6 and 24 months in 2011, wherein the tuple representation of the active power flow in four days and the active power flow pair in 9 and 24 days are shown in fig. 6a to 6 d. As can be seen from fig. 6a and fig. 6d, the active power flow data of day 12 in month 6 and day 24 in month 9 are the closest (the shortest distance is), so the active power flow data of day 13 in month 6 is used to predict the active power flow at each moment in day 25 in month 9. Fig. 7 is a graph comparing the predicted result with the actual power flow of 9 months and 25 days. As can be seen from fig. 7, although there are some differences in the values of the actual values and the predicted values, the overall trend is correct and the shapes are also very similar. In addition, the method can be applied to real-time calculation occasions, so that the latest prediction result can be given according to the latest data by continuous correction.
Analysis of the above examples shows that: the technical scheme of the invention overcomes the defects of the existing prediction method based on meteorological prediction in the aspect of cluster point wind power prediction, compared with the existing prediction method based on historical data, the prediction method based on meteorological prediction adopts piecewise linearity to express tuple sequence, extracts the main characteristics of the historical data, has the characteristics of high search speed, high accuracy and the like, and also provides a transformation coefficient to correct the historical data by considering that the phenomena generated under the same mechanism are not completely consistent but are similar so as to obtain a more accurate prediction result.
The SWAB algorithm is a fast robust pattern matching algorithm applied in time series databases in the article "A fast and robust method for pattern matching in time series databases". Paper author KEOGH e, paper name chinese translation "a fast robust pattern matching algorithm applied in time series databases" [ C ] ninth international association of artificial intelligence tools proceedings, new harbor beach city: IEEE (or institute of Electrical and electronics Engineers, USA), 1997: 578-.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (5)
1. A cluster point wind power prediction method based on tuple matching is characterized by comprising the following steps:
step 1: inputting active power data of all historical wind power outputs of the wind power cluster points;
step 2: carrying out piecewise linearization processing on wind power historical output data of the wind power cluster points by utilizing a piecewise linearization method, and expressing a processed result by adopting a tuple sequence;
and step 3: determining a sequence S of tuples to be matched0And extracting the tuple sequence formed in the step 2Sequence set S of row-component tuplessetThe tuple sequence S to be matched0Is the current time T1Push forward a complete day d, time interval [ T1-d,T1]The tuple sequence of the internal wind power output data;
and 4, step 4: solving the tuple sequence set S by using a tuple dynamic distortion matching algorithmsetNeutralizing tuple sequence S to be matched0Best matched target tuple sequence S1(ii) a The tuple dynamic distortion matching algorithm is as follows:
the tuple dynamic distortion matching algorithm is a nonlinear warping algorithm combining time warping, tuple representation and distance measurement calculation, a mapping function phi () is found, and a tuple is testedIs nonlinearly mapped to the reference tupleAnd such that the function satisfies:
wherein,are respectively a test tuple sequence SaAnd a reference tuple sequence SbThe number of n, m tuples in (a),is the distance between two tuples, D is the distance between two tuple sequences under the optimal time warping condition, N is the tuple sequence SaThe number of tuples in (1);
tuple metric function of an algorithmIs calculated as follows：
Is provided with a tuple sequence Sa，Sb，Andare respectively Sa,SbN, m segmented tuples of (S)a,SbRespectively has a sequence length of delta A, delta B, w0For the weighting coefficients, the tuple distance calculation function is as follows
Wherein the weighting coefficient function:
magnitude distance function:
andexpressed as a sequence S of test tuplesaAnd a reference tuple sequence SbThe slope of (a) of (b) is,
the cumulative distance of the minimum path is derived using a cumulative distance calculation formula as follows:
d is the distance of the two tuple sequences under the optimal time warping condition;
the tuple sequence refers to a method for representing piecewise linearized data, the tuple refers to a set of discrete data, and a line segment is represented by using a 4-tuple L (x, y, k, delta x), wherein x is the abscissa of any point on a straight line, y is the ordinate of the point, k is the slope of the straight line, and delta x is the projection length of the line segment on the x axis; the tuple sequence is an ordered set of tuples, and for the history data after m segments, the tuple sequence is expressed as:
S＝{L1(x1,y1,k1,Δx1),…,Lm(xm,ym,km,Δxm)}；
and 5: calculating a tuple sequence S to be matched0And a sequence of target tuples S1The transform coefficients between; the transformation coefficient refers to the time compression ratio lambdatimeTime offset amount Δ xtimeAnd amplitude compression ratio lambdapowerThe method comprises the following steps:
obtaining a target tuple sequence S according to the matching path1And tuple sequence S to be matched0The one-to-one correspondence of the tuples in the system is provided with n tuple pairs Andsequence of tuples S1,S2I, j tuples of (1);
the time compression ratio lambdatime：
The time compression rate is characterized by the average compression degree between the two tuple sequences on a time axis;
time offset amount Δ xtime
The time offset represents the dislocation degree of the two tuple sequences on the time axis;
amplitude compression ratio lambdapower
The amplitude compression rate is characterized by the average compression degree of the two-tuple sequence on the amplitude
Step 6: utilizing the transformation coefficient obtained in the step 5 to pair the target tuple sequence S1Calculating the subsequent historical data; and the operation of the transformation coefficient on the historical data is as follows: after every two adjacent data points of the historical data are connected, the function of the historical data is expressed as H (t), and if t is more than or equal to 0, the operation result is as follows:
H′(t)＝λpower·H(λtime·(t-Δxtime)),t≥0；
and 7: and outputting the predicted value of the wind power of the cluster point.
2. The tuple matching-based cluster point wind power prediction method according to claim 1, wherein the piecewise linearization method is a hybrid piecewise linearization algorithm adopting a SWAB algorithm, the SWAB algorithm is a combination of a classic slipping-Window algorithm and a Bottom-Up algorithm, the buffering area is dynamically updated by the slipping-Window algorithm, and the Bottom-Up algorithm performs piecewise linearization again on data in the buffering area.
3. According to the rightThe method for predicting the wind power of the cluster points based on the tuple matching as claimed in claim 1, wherein the tuple sequence set S issetIn the historical wind power output data of cluster points, the time periods are respectively
[T1-2d,T1-d],[T1-3d,T1-2d],…,[T1-nd,T1-(n-1)d]…, the data are respectively processed by segment linearization and tuple representation to form tuple sequence set.
4. The tuple matching-based wind power prediction method for cluster points according to claim 1, wherein the historical data after the target tuple sequence is when the time interval of the target tuple sequence is [ T ]1,T2]T in time and cluster point wind power output historical data2All data after the time.
5. The tuple matching-based cluster point wind power prediction method according to claim 4, wherein the outputting of the cluster point wind power prediction value is:
discretizing the H ' (T) at the sampling time interval DeltaT of the original historical data, and taking data points H ' (0), H ' (1), …, H ' (T/DeltaT) with a required time length T from the discretized H ' (n) as prediction data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310451220.1A CN103577890B (en) | 2013-09-25 | 2013-09-25 | Based on the matched cluster point wind power forecasting method of tuple |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310451220.1A CN103577890B (en) | 2013-09-25 | 2013-09-25 | Based on the matched cluster point wind power forecasting method of tuple |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103577890A CN103577890A (en) | 2014-02-12 |
| CN103577890B true CN103577890B (en) | 2019-01-29 |
Family
ID=50049633
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310451220.1A Expired - Fee Related CN103577890B (en) | 2013-09-25 | 2013-09-25 | Based on the matched cluster point wind power forecasting method of tuple |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103577890B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104318484B (en) * | 2014-09-30 | 2018-07-27 | 东南大学 | A kind of cluster wind power plant volatility modeling method |
| CN104751241B (en) * | 2015-03-25 | 2017-10-31 | 广西大学 | A kind of method of wind power interval prediction |
| US10168449B2 (en) | 2016-08-15 | 2019-01-01 | International Business Machines Corporation | Correcting computer model weather forecasts using a hybrid analog method with dynamic time warping |
| CN112906928B (en) * | 2019-12-03 | 2022-09-16 | 国网山西省电力公司电力科学研究院 | Wind power plant cluster active power prediction method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102682185A (en) * | 2011-03-10 | 2012-09-19 | 华锐风电科技(集团)股份有限公司 | Single wind turbine wind power prediction method |
| CN103246702A (en) * | 2013-04-02 | 2013-08-14 | 大连理工大学 | Industrial sequential data missing filling method based on sectional state displaying |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013106941A1 (en) * | 2012-01-17 | 2013-07-25 | Energiebüro AG | Method of predicting the expected future power production of a wind or solar power plant |
-
2013
- 2013-09-25 CN CN201310451220.1A patent/CN103577890B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102682185A (en) * | 2011-03-10 | 2012-09-19 | 华锐风电科技(集团)股份有限公司 | Single wind turbine wind power prediction method |
| CN103246702A (en) * | 2013-04-02 | 2013-08-14 | 大连理工大学 | Industrial sequential data missing filling method based on sectional state displaying |
Non-Patent Citations (2)
| Title |
|---|
| 基于多目标优化的含风电场的电力系统经济调度;常鹏;《中国优秀硕士学位论文全文数据库 工程科技II辑》;20130615(第6期);第1-4章 |
| 基于大规模风电集中接入电网的无功电压特性及控制研究;梁纪峰;《中国优秀硕士学位论文全文数据库 工程科技II辑》;20130115(第1期);第2-3章 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103577890A (en) | 2014-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kong et al. | Wind speed prediction using reduced support vector machines with feature selection | |
| Dong et al. | Wind power day-ahead prediction with cluster analysis of NWP | |
| Wang et al. | Research and application of a combined model based on multi-objective optimization for multi-step ahead wind speed forecasting | |
| Zjavka | Wind speed forecast correction models using polynomial neural networks | |
| Zhang et al. | A hybrid prediction model for forecasting wind energy resources | |
| Sun et al. | Staged icing forecasting of power transmission lines based on icing cycle and improved extreme learning machine | |
| CN103577890B (en) | Based on the matched cluster point wind power forecasting method of tuple | |
| Wang et al. | An innovative hybrid approach for multi-step ahead wind speed prediction | |
| CN106485368A (en) | A kind of short term prediction method for energy resource system power output | |
| CN103985000B (en) | Medium-and-long term typical daily load curve prediction method based on function type nonparametric regression | |
| CN103401236A (en) | Wind power farm generator unit grouping method based on flow correlation of wind power farm | |
| CN109242169B (en) | Short-term load prediction method for large power customer | |
| CN111027775A (en) | Step hydropower station generating capacity prediction method based on long-term and short-term memory network | |
| CN106600037B (en) | Multi-parameter auxiliary load prediction method based on principal component analysis | |
| CN103020743B (en) | Wind energy turbine set ultra-short term wind speed forecasting method | |
| CN104036356A (en) | Method for predicting future operating state of power grid by using fractal algorithm | |
| CN114021483A (en) | Ultra-short-term wind power prediction method based on time domain characteristics and XGboost | |
| Liu et al. | An online transfer learning model for wind turbine power prediction based on spatial feature construction and system-wide update | |
| Chu et al. | NWP combination correction model based on variable-weight stacking algorithm | |
| CN107579516B (en) | Method for improving state estimation calculation speed of power system | |
| CN103489037A (en) | Forecasting method of wind resource capable of generating power | |
| Li et al. | A novel discrete grey model with time-delayed power term and its application in solar power generation volume forecasting | |
| Kuang et al. | Short-term power load forecasting method in rural areas based on cnn-lstm | |
| Li | The Study of Short Term Wind Power Prediction Based on MV-LSTM | |
| Lijie et al. | Short-term wind power prediction with signal decomposition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190129 Termination date: 20200925 |