CN111680823A - Wind direction information prediction method and system - Google Patents
Wind direction information prediction method and system Download PDFInfo
- Publication number
- CN111680823A CN111680823A CN202010397191.5A CN202010397191A CN111680823A CN 111680823 A CN111680823 A CN 111680823A CN 202010397191 A CN202010397191 A CN 202010397191A CN 111680823 A CN111680823 A CN 111680823A
- Authority
- CN
- China
- Prior art keywords
- wind direction
- prediction
- unit
- time
- historical wind
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000005070 sampling Methods 0.000 claims abstract description 47
- 238000012216 screening Methods 0.000 claims abstract description 21
- 238000012545 processing Methods 0.000 claims description 30
- 238000012952 Resampling Methods 0.000 claims description 19
- 238000004064 recycling Methods 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000013075 data extraction Methods 0.000 claims description 4
- 239000000284 extract Substances 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims 1
- 230000008447 perception Effects 0.000 abstract description 3
- 238000011217 control strategy Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 230000002411 adverse Effects 0.000 description 4
- 238000012937 correction Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000010219 correlation analysis Methods 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000003334 potential effect Effects 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a wind direction information prediction method and a system, wherein the method comprises the following steps: s1: acquiring historical wind direction real-time data of each wind turbine of a wind power plant; s2: screening a plurality of associated units of a target unit according to the historical wind direction real-time data; s3: extracting historical wind direction real-time data of a target unit and an associated unit at p sampling moments before the current prediction moment to form a historical wind direction time sequence of the target unit and a historical wind direction time sequence of the associated unit, wherein p is a constant; s4: inputting the historical wind direction time sequence of the target unit and the historical wind direction time sequence of the associated unit into a pre-established prediction model to obtain a wind direction prediction value of the target unit at the current prediction time; s5: when the next prediction time is reached, the historical wind direction real-time data in step S1 is updated, and steps S2 to S5 are repeatedly executed until the prediction is completed. Compared with the prior art, the method disclosed by the invention integrates the time sequence rule and the space related information, and improves the accuracy and stability of wind direction perception.
Description
Technical Field
The invention belongs to the field of wind power, and particularly relates to a wind direction information prediction method and system.
Background
The wind direction has obvious randomness and volatility, the traditional wind turbine yaw control strategy has the problems of action lag, high yaw frequency and the like, and the adverse effects on the wind energy utilization efficiency and the reliability of a fan yaw system can be caused. The yaw control strategy based on wind direction prediction can effectively improve wind energy capturing efficiency and improve the safety and economy of wind turbine generator operation. The wind direction prediction method for the wind turbine generator yaw control can improve the lagging yaw control of the large wind turbine generator, effectively optimize the working performance of a yaw system, reduce frequent actions of a yaw mechanism caused by wind direction fluctuation, reduce unbalanced load of a wind wheel, and meet the internal requirement of ensuring safe and stable operation of the wind turbine generator.
At present, certain achievements have been obtained in the field of wind parameter prediction, and from the viewpoint of prediction principles, the wind parameter prediction mainly includes a physical model method, a statistical method, a machine learning method and the like. Under the influence of factors such as air pressure, terrain, turbulence and the like, wind direction prediction faces the challenges of complexity, variability and weak regularity. The existing prediction method is used for establishing a time sequence prediction model based on an internal time sequence rule of historical data or establishing a machine learning model based on input variables such as terrain, air pressure, temperature, relative humidity, geographic position, altitude and the like, and the common point of the existing prediction method is that the existing prediction method is limited to data of an independent unit or a wind power plant for prediction and interactive utilization of space information is lacked, so that when a wind speed/wind direction sequence is not stable, the method has a large prediction error and can cause serious adverse effects on yaw control based on wind direction prediction, the accuracy of the wind direction prediction needs to be concerned, and the stability of the wind direction prediction needs to be improved.
Disclosure of Invention
The present invention is directed to a method and a system for predicting wind direction information, which overcome the above-mentioned drawbacks of the prior art.
The purpose of the invention can be realized by the following technical scheme:
a wind direction information prediction method comprises the following steps:
s1: acquiring historical wind direction real-time data of each wind turbine of a wind power plant;
s2: screening a plurality of associated units of a target unit according to the historical wind direction real-time data;
s3: extracting historical wind direction real-time data of a target unit and an associated unit at p sampling moments before the current prediction moment to form a historical wind direction time sequence of the target unit and a historical wind direction time sequence of the associated unit, wherein p is a constant;
s4: inputting the historical wind direction time sequence of the target unit and the historical wind direction time sequence of the associated unit into a pre-established prediction model to obtain a wind direction prediction value of the target unit at the current prediction time;
s5: when the next prediction time is reached, the historical wind direction real-time data in step S1 is updated, and steps S2 to S5 are repeatedly executed until the prediction is completed.
Preferably, the step S2 includes preprocessing the real-time data of the historical wind direction before screening the associated units, specifically:
firstly, a recycling process:
let the original dataset be Xr＝{x1,x2,…,xnObtaining a recycling processing data set X 'after recycling processing'r＝{x′1,x′2,…,x′nAnd (c) the step of (c) in which,
x′1＝x1，
when the value of i is greater than 1,
wherein x isiFor the ith historical wind direction real-time data, xiThe data is ith historical wind direction real-time data after the circulation processing, wherein i is 1, 2, … … n, and n is the total number of data in an original data set;
then, resampling:
setting a resampling interval as T seconds, wherein each resampling interval corresponds to d data in the data set for de-circulation processing, and sequentially dividing each d data in the data set for de-circulation processing into a group and calculating an average value to obtain the real-time data of the re-sampled historical wind direction;
and then, screening the associated units based on the resampled historical wind direction real-time data.
Preferably, the prediction model is a vector autoregressive prediction model.
Preferably, the vector autoregressive prediction model is expressed in the form of:
setting a historical wind direction time sequence of the target unit as follows: x ═ Xt-p,xt-p+1,…,xt-2,xt-1}，
The 1 st associated unit historical wind direction time sequence is as follows: y ═ Yt-p,yt-p+1,…,yt-2,yt-1}，
By analogy, the historical wind direction time sequence of the mth associated unit is as follows: z ═ Zt-p,zt-p+1,…,zt-2,zt-1}，
Wherein x ist-jHistorical wind direction real-time data y of a target unit at the jth sampling moment before the current prediction momentt-jHistorical wind direction real-time data z of the 1 st associated unit at the jth sampling moment before the current prediction momentt-jThe historical wind direction real-time data of the mth associated unit at the jth sampling time before the current prediction time is obtained, j is 1, 2, … … p, and m is the total number of the associated units;
the vector autoregressive prediction model is then:
wherein x istFor the predicted value of the wind direction of the target unit at the current prediction time, w0Is a constant value, w0Representing the stationary terms of the historical wind direction time series of the target unit, αjβ co-order co-integration influence factor of the historical wind direction real-time data of the j sampling moment before the current prediction moment of the target unit per se on the wind direction prediction value of the target unit at the current prediction momentjRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the 1 st associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment, and so on, gammajRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the mth associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment,tis a random perturbation term with a mean of 0 and a constant variance, αj、βj、γjAndtare all constants obtained by pre-training.
Preferably, the method is used for yaw control of the target unit after acquiring the wind direction predicted value of the target unit at the current prediction time.
A wind direction information prediction system, the system comprising:
a data acquisition unit: the unit is used for acquiring historical wind direction real-time data of each wind turbine in the wind power plant;
and a related unit screening unit: the unit screens a plurality of associated units of a target unit according to historical wind direction real-time data;
a data extraction unit: the unit extracts historical wind direction real-time data of a target unit and an associated unit at p sampling moments before the current prediction moment to form a historical wind direction time sequence of the target unit and a historical wind direction time sequence of the associated unit, wherein p is a constant;
a prediction unit: the unit inputs the historical wind direction time sequence of the target unit and the historical wind direction time sequence of the associated unit into a pre-established prediction model to obtain a wind direction prediction value of the target unit at the current prediction time.
Preferably, the associated unit screening unit includes:
a pretreatment subunit: the sub-unit performs the following operations,
firstly, a recycling process:
let the original dataset be Xr＝{x1,x2,…,xnObtaining a recycling processing data set X 'after recycling processing'r＝{x′1,x′2,…,x′nAnd (c) the step of (c) in which,
x′1＝x1，
when the value of i is greater than 1,
wherein x isiFor the ith historical wind direction real-time data, xi' is the ith historical wind direction real-time data after the circulation removing processing, i is 1,2, … … n, n is the total number of data in the original data set;
then, resampling:
setting a resampling interval as T seconds, wherein each resampling interval corresponds to d data in the data set for de-circulation processing, and sequentially dividing each d data in the data set for de-circulation processing into a group and calculating an average value to obtain the real-time data of the re-sampled historical wind direction;
screening subunits: the subunit screens the associated units based on the resampled historical wind direction real-time data.
Preferably, the prediction model in the prediction unit is a vector autoregressive prediction model.
Preferably, the vector autoregressive prediction model is expressed in the form of:
setting a historical wind direction time sequence of the target unit as follows: x ═ Xt-p,xt-p+1,…,xt-2,xt-1}，
The 1 st associated unit historical wind direction time sequence is as follows: y ═ Yt-p,yt-p+1,…,yt-2,yt-1}，
By analogy, the historical wind direction time sequence of the mth associated unit is as follows: z ═ Zt-p,zt-p+1,…,zt-2,zt-1}，
Wherein x ist-jHistorical wind direction real-time data y of a target unit at the jth sampling moment before the current prediction momentt-jHistorical wind direction real-time data z of the 1 st associated unit at the jth sampling moment before the current prediction momentt-jThe historical wind direction real-time data of the mth associated unit at the jth sampling time before the current prediction time is obtained, j is 1, 2, … … p, and m is the total number of the associated units;
the vector autoregressive prediction model is then:
wherein x istFor the predicted value of the wind direction of the target unit at the current prediction time, w0Is a constant value, w0Representing the stationary terms of the historical wind direction time series of the target unit, αjβ co-order co-integration influence factor of the historical wind direction real-time data of the j sampling moment before the current prediction moment of the target unit per se on the wind direction prediction value of the target unit at the current prediction momentjRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the 1 st associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment, and so on, gammajRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the mth associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment,tis a random perturbation term with a mean of 0 and a constant variance, αj、βj、γjAndtare all constants obtained by pre-training.
Preferably, the system is connected with a fan yaw control system, and the prediction unit is used for yaw control of a target unit in the fan yaw control system after obtaining a wind direction prediction value of the target unit at the current prediction time.
Compared with the prior art, the invention has the following advantages:
(1) according to the method and the system, the wind direction data time sequence characteristics and the space correlation characteristics of the wind generation sets of the wind power plant are combined, the time sequence rule and the space correlation information are fused, and the accuracy and the stability of wind direction perception are improved;
(2) the comprehensive utilization degree of the time sequence rule and the spatial correlation information of the vector autoregressive prediction model is high, and prediction close to reality can be provided during prediction;
(3) the method and the system are used for controlling the yaw of the fan, are favorable for improving the utilization efficiency of wind energy and prolonging the service life of a yaw mechanism.
Drawings
FIG. 1 is a flow chart of a wind direction information prediction method according to the present invention;
FIG. 2 is a block diagram illustrating a wind direction information prediction method according to the present invention;
FIG. 3 is a schematic view of a wind direction data cycle according to the present invention;
FIG. 4 is a schematic diagram of a vector autoregressive model according to the present invention;
FIG. 5 is a block diagram of a flow of wind direction prediction using the wind direction information prediction method of the present invention;
FIG. 6 is a schematic structural diagram of a wind direction information prediction system according to the present invention;
in the figure, 1 is a data acquisition unit, 2 is a related unit screening unit, 3 is a data extraction unit, 4 is a prediction unit, 21 is a preprocessing subunit, and 22 is a screening subunit.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. Note that the following description of the embodiments is merely a substantial example, and the present invention is not intended to be limited to the application or the use thereof, and is not limited to the following embodiments.
Examples
As shown in fig. 1, a wind direction information prediction method includes the following steps:
s1: acquiring historical wind direction real-time data of each wind turbine of a wind power plant;
s2: screening a plurality of associated units of a target unit according to the historical wind direction real-time data;
s3: extracting historical wind direction real-time data of a target unit and an associated unit at p sampling moments before the current prediction moment to form a historical wind direction time sequence of the target unit and a historical wind direction time sequence of the associated unit, wherein p is a constant;
s4: inputting the historical wind direction time sequence of the target unit and the historical wind direction time sequence of the associated unit into a pre-established prediction model to obtain a wind direction prediction value of the target unit at the current prediction time;
s5: when the next prediction time is reached, the historical wind direction real-time data in step S1 is updated, and steps S2 to S5 are repeatedly executed until the prediction is completed.
FIG. 2 is a prediction block diagram of the wind direction information prediction method, according to the method, wind direction data time sequence characteristics and space correlation characteristics of wind power plants and wind power generation sets of a wind power plant are combined, time sequence rules and space correlation information are fused, and accuracy and stability of wind direction perception are improved, wherein the space correlation characteristics refer to that correlation sets are screened for prediction of wind directions of target sets, and the data time sequence characteristics refer to that historical wind direction time sequence sequences of the target sets and the correlation sets are used for prediction of wind directions of the target sets. As can be seen from the figure, a plurality of time sequences need to be input at the input end of the prediction model at the same time, and the prediction process needs to take account of the ambiguity of the time sequence rule and the variability of time-lag relations between different sequences. In order to improve the accuracy of the prediction model, the deviation between the prediction result and the actual value is finally returned to the prediction model as a parameter for feedback verification and adjustment. Specifically, the error superposition correction needs to be performed on the prediction model according to the historical prediction result and the actual data. And error superposition correction, namely, firstly, obtaining an original result of wind direction prediction by adopting an arbitrary prediction method, then, constructing an error prediction model, performing full training under reasonable iteration times, predicting errors contained in a predicted value, superposing the errors with the original predicted value, and taking a superposed value as a final prediction result.
According to the actual value x of the wind direction sequencereal,tAnd sequence predictor xpred,tAnd obtaining a prediction error sequence:
et＝xreal,t-xpred,t
the error correction of the invention adopts the simplest Engle-Granger two-step method to carry out feedback check on the prediction error.
The wind direction prediction method is oriented to unit control application, and the prediction method is verified by data with sampling intervals of minute levels, so that original data needs to be resampled. In order to avoid the problem that the accuracy of the resampled data is greatly influenced due to the abnormal data recording of a certain point in the original data, the data at the sampling point needs to be averaged according to the principle of data recording of the SCADA system. However, the wind direction data is recorded by angle data and has cyclicity, as shown in fig. 3, when the wind direction data simultaneously contains the I, IV th quadrant, directly calculating the average value may cause some problems and cause statistical errors of inter-unit correlation, and therefore, the step S2 includes preprocessing of the real-time data of the historical wind direction before screening the associated units, specifically:
firstly, a recycling process:
let the original dataset be Xr＝{x1,x2,…,xnObtaining a recycling processing data set X 'after recycling processing'r＝{x′1,x′2,…,x′nAnd (c) the step of (c) in which,
x′1＝x1，
when the value of i is greater than 1,
wherein x isiFor the ith historical wind direction real-time data, xiThe data is ith historical wind direction real-time data after the circulation processing, wherein i is 1, 2, … … n, and n is the total number of data in an original data set;
then, resampling:
setting a resampling interval as T seconds, wherein each resampling interval corresponds to d data in the data set for de-circulation processing, and sequentially dividing each d data in the data set for de-circulation processing into a group and calculating an average value to obtain the real-time data of the re-sampled historical wind direction;
furthermore, the relevant units are screened based on the resampled historical wind direction real-time data, and a method for screening the relevant units is disclosed in chinese patent CN110296055A, and is not described in detail herein.
The prediction model of the invention is a vector autoregressive prediction model, and specifically, as shown in fig. 4, the mathematical expression form of the vector autoregressive prediction model is as follows:
setting a historical wind direction time sequence of the target unit as follows: x ═ Xt-p,xt-p+1,…,xt-2,xt-1}，
The 1 st associated unit historical wind direction time sequence is as follows: y ═ Yt-p,yt-p+1,…,yt-2,yt-1}，
By analogy, the historical wind direction time sequence of the mth associated unit is as follows: z ═ Zt-p,zt-p+1,…,zt-2,zt-1}，
Wherein x ist-jHistorical wind direction real-time data y of a target unit at the jth sampling moment before the current prediction momentt-jHistorical wind direction real-time data z of the 1 st associated unit at the jth sampling moment before the current prediction momentt-jThe historical wind direction real-time data of the mth associated unit at the jth sampling time before the current prediction time is obtained, j is 1, 2, … … p, and m is the total number of the associated units;
the vector autoregressive prediction model is then:
wherein x istFor the predicted value of the wind direction of the target unit at the current prediction time, w0Is a constant value, w0Representing the stationary terms of the historical wind direction time series of the target unit, αjβ co-order co-integration influence factor of the historical wind direction real-time data of the j sampling moment before the current prediction moment of the target unit per se on the wind direction prediction value of the target unit at the current prediction momentjRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the 1 st associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment, and so on, gammajRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the mth associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment,tis a random perturbation term with a mean of 0 and a constant variance, αj、βj、γjAndtare all constants obtained by pre-training.
As shown in fig. 5, a flow chart of the wind direction prediction by applying the method of the present invention is shown in fig. 4: in the figure, the x-axis is a time axis and represents the state of the target unit i at different moments, and the y-axis represents the extension of the prediction flow at each moment. At t0Screening the associated units of i at any moment, extracting the spatial correlation characteristics and the time sequence rule characteristics of the associated units, inputting the spatial correlation characteristics and the time sequence rule characteristics into a prediction model established by a machine learning method, and obtaining t0+1Predicted value of time(ii) a To t0+1At the moment, the space correlation characteristics and the time sequence regular characteristics of the target unit and the related unit are extracted and input into a prediction model to predict t0+2The wind direction value at the moment, and at the same time, t0+1Actual value of time and t0T obtained at time0+1And comparing the predicted values at the moment, and feeding the error sequence back to the prediction model by using the error correction model as a check parameter to realize feedback adjustment of the model.
The method is used for yaw control of the target unit after the wind direction predicted value of the target unit at the current prediction time is obtained.
As shown in fig. 6, a wind direction information prediction system includes:
the data acquisition unit 1: the unit is used for acquiring historical wind direction real-time data of each wind turbine in the wind power plant;
associated unit screening unit 2: the unit screens a plurality of associated units of a target unit according to historical wind direction real-time data;
the data extraction unit 3: the unit extracts historical wind direction real-time data of a target unit and an associated unit at p sampling moments before the current prediction moment to form a historical wind direction time sequence of the target unit and a historical wind direction time sequence of the associated unit, wherein p is a constant;
the prediction unit 4: the unit inputs the historical wind direction time sequence of the target unit and the historical wind direction time sequence of the associated unit into a pre-established prediction model to obtain a wind direction prediction value of the target unit at the current prediction time.
The associated unit screening unit 2 includes:
the preprocessing subunit 21: the sub-unit performs the following operations,
firstly, a recycling process:
let the original dataset be Xr＝{x1,x2,…,xnObtaining a recycling processing data set X 'after recycling processing'r＝{x′1,x′2,…,x′nAnd (c) the step of (c) in which,
x′1＝x1，
when the value of i is greater than 1,
wherein x isiFor the ith historical wind direction real-time data, xiThe data is ith historical wind direction real-time data after the circulation processing, wherein i is 1, 2, … … n, and n is the total number of data in an original data set;
then, resampling:
setting a resampling interval as T seconds, wherein each resampling interval corresponds to d data in the data set for de-circulation processing, and sequentially dividing each d data in the data set for de-circulation processing into a group and calculating an average value to obtain the real-time data of the re-sampled historical wind direction;
screening subunit 22: the subunit screens the associated units based on the resampled historical wind direction real-time data.
The prediction model in the prediction unit 4 is a vector autoregressive prediction model.
The vector autoregressive prediction model is expressed in the form of mathematical expression:
setting a historical wind direction time sequence of the target unit as follows: x ═ Xt-p,xt-p+1,…,xt-2,xt-1}，
The 1 st associated unit historical wind direction time sequence is as follows: y ═ Yt-p,yt-p+1,…,yt-2,yt-1}，
By analogy, the historical wind direction time sequence of the mth associated unit is as follows: z ═ Zt-p,zt-p+1,…,zt-2,zt-1}，
Wherein x ist-jHistorical wind direction real-time data y of a target unit at the jth sampling moment before the current prediction momentt-jHistorical wind direction real-time data z of the 1 st associated unit at the jth sampling moment before the current prediction momentt-jThe historical wind direction real-time data of the mth associated unit at the jth sampling time before the current prediction time is obtained, j is 1, 2, … … p, and m is the total number of the associated units;
the vector autoregressive prediction model is then:
wherein x istFor the predicted value of the wind direction of the target unit at the current prediction time, w0Is a constant value, w0Representing the stationary terms of the historical wind direction time series of the target unit, αjβ co-order co-integration influence factor of the historical wind direction real-time data of the j sampling moment before the current prediction moment of the target unit per se on the wind direction prediction value of the target unit at the current prediction momentjRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the 1 st associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment, and so on, gammajRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the mth associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment,tis a random perturbation term with a mean of 0 and a constant variance, αj、βj、γjAndtare all constants obtained by pre-training.
The system is connected with a fan yaw control system, and a prediction unit 4 is used for yaw control of a target unit in the fan yaw control system after obtaining a wind direction predicted value of the target unit at the current prediction time.
Evaluation indexes of the wind direction prediction method:
i) accuracy of measurement
One common measure of prediction is accuracy, which refers to the deviation between the predicted value of wind direction and the actual value. The wind direction prediction method is influenced by factors such as air pressure, terrain, turbulence and the like, the wind direction prediction faces the challenges of complexity, changeability and weak regularity, the accuracy of the existing wind direction prediction method needs to be further improved, and the guiding significance to the prediction control of the wind turbine yaw system is limited. On the one hand, because the wind direction is a directional vector, the prediction precision of the wind direction relates to the magnitude and the positive and negative of the yaw error angle, the magnitude of the prediction error determines whether the yaw controller gives a command or not, and the positive and negative of the prediction error determines the alignment error of the yaw direction. If the prediction error is a negative value, the yaw direction is opposite to the actual demand direction of the wind turbine generator, so that the wind energy capturing efficiency of the wind turbine generator is reduced, the invalid yaw frequency is increased, the potential action cost is increased, and the service life of a yaw system is adversely affected. On the other hand, due to the limitations of wind direction measurement techniques, there is a certain error between the measured value of the wind direction itself and the actual value, and if the prediction error cannot be reduced as much as possible, the overall error between the data transmitted to the yaw controller and the actual wind direction further increases.
The prediction accuracy is evaluated using the mean absolute error MAE and the root mean square error RMSE:
wherein y isi,i∈[1,n]The actual value is represented by the value of,
indicating the predicted value.
ii) stability
In the wind direction prediction for control application, if a large error occurs to cause misoperation, the economic benefit of predicting yaw is reduced, so that the requirement on the stability of prediction is high. The predicted stability refers to the degree of dispersion between the predicted result and the actual result, the level of which depends on the stability of the prediction model structure. If the prediction error is easy to have a large value, the stability is low. The existing prediction method is based on the internal time sequence rule of historical data or is based on input variables such as terrain, air pressure, temperature, relative humidity, geographical position, altitude and the like to establish a machine learning model, the common points of the existing prediction method are that the existing prediction method is limited in data of independent units or wind power plants to carry out prediction, interactive utilization of information among the units is lacked, and when a wind direction sequence is not stable, a large prediction error is often generated, so that error control operation is caused, and serious adverse effects are caused. Therefore, not only the accuracy of prediction needs to be concerned, but also the stability of prediction needs to be improved when a wind direction prediction model is established.
The prediction stability can be evaluated by using a coefficient of variation, which is more sensitive to larger errors occurring in the prediction, so that the coefficient of variation can be used as a natural measure of the prediction stability:
iii) time scale
According to the working principle of the wind turbine yaw system, in the wind turbine yaw control strategy optimization, the wind direction prediction time scale has great influence on the wind energy capturing efficiency and the yaw frequency. The wind direction prediction time scale is directly related to the wind direction prediction precision and the feasibility of using the wind direction prediction result in an actual yaw control strategy. When the prediction time length is determined, the yaw delay Ty needs to be considered, the restart wind yaw delay is related to the rated power of the unit, the advance of wind direction prediction needs to cover the yaw delay, and otherwise, the advance has no guiding significance on yaw. Defining the minimum prediction time length as Tmin, then
Tmin≥Tymax
The time delay Ty in the yaw control strategy is set manually and is generally between 5 and 120 s.
In the embodiment, the data source is actual operation data of a certain wind power plant in Zhan province, in order to verify the effectiveness and universality of the prediction method, 1min sampling data (group A) and 10min sampling data (group B) are respectively adopted to verify the prediction model, and in data resampling, the group A data is a 5-second wind direction average value at an interval of 1 minute, namely T dataA＝60s，dAGroup B data is the average value of 25 second wind direction at 5 minute intervals, namely TB＝300s，dB＝25。
In A, B two sets of data, two sets were randomly selected as the predicted target set, and then 3 spatially correlated sets were selected based on the correlation analysis of the yaw events. According to the results of random selection and cross-correlation analysis, the target unit is predicted to be a 30# unit from the group A data, and the selected associated units are 2#, 17# and 18# units; and predicting the target unit to be a 24# unit from the B group data, and selecting the associated units to be 2#, 14# and 20# units. The two groups of data prediction result evaluations are shown in table 1, so that the method has a good prediction effect, and meanwhile, the prediction results also verify that the wind direction prediction by fully utilizing the spatial correlation has good technical economy and technical feasibility.
TABLE 1A, B predictive evaluation of the two sets of data
The above embodiments are merely examples and do not limit the scope of the present invention. These embodiments may be implemented in other various manners, and various omissions, substitutions, and changes may be made without departing from the technical spirit of the present invention.
Claims (10)
1. A wind direction information prediction method is characterized by comprising the following steps:
s1: acquiring historical wind direction real-time data of each wind turbine of a wind power plant;
s2: screening a plurality of associated units of a target unit according to the historical wind direction real-time data;
s3: extracting historical wind direction real-time data of a target unit and an associated unit at p sampling moments before the current prediction moment to form a historical wind direction time sequence of the target unit and a historical wind direction time sequence of the associated unit, wherein p is a constant;
s4: inputting the historical wind direction time sequence of the target unit and the historical wind direction time sequence of the associated unit into a pre-established prediction model to obtain a wind direction prediction value of the target unit at the current prediction time;
s5: when the next prediction time is reached, the historical wind direction real-time data in step S1 is updated, and steps S2 to S5 are repeatedly executed until the prediction is completed.
2. The wind direction information prediction method according to claim 1, wherein the step S2 includes preprocessing the real-time data of the historical wind direction before screening the associated units, specifically:
firstly, a recycling process:
let the original dataset be Xr＝{x1,x2,…,xnObtaining a recycling processing data set X 'after recycling processing'r＝{x′1,x′2,…,x′nAnd (c) the step of (c) in which,
x′1＝x1，
when the value of i is greater than 1,
wherein x isiFor the ith historical wind direction real-time data, xiThe data is ith historical wind direction real-time data after the circulation processing, wherein i is 1, 2, … … n, and n is the total number of data in an original data set;
then, resampling:
setting a resampling interval as T seconds, wherein each resampling interval corresponds to d data in the data set for de-circulation processing, and sequentially dividing each d data in the data set for de-circulation processing into a group and calculating an average value to obtain the real-time data of the re-sampled historical wind direction;
and then, screening the associated units based on the resampled historical wind direction real-time data.
3. The method of claim 1, wherein the prediction model is a vector autoregressive prediction model.
4. The method according to claim 3, wherein the vector autoregressive prediction model is expressed mathematically as:
setting a historical wind direction time sequence of the target unit as follows: x ═ Xt-p,xt-p+1,…,xt-2,xt-1}，
The 1 st associated unit historical wind direction time sequence is as follows: y ═ Yt-p,yt-p+1,…,yt-2,yt-1}，
By analogy, the historical wind direction time sequence of the mth associated unit is as follows: z ═ Zt-p,zt-p+1,…,zt-2,zt-1}，
Wherein x ist-jHistorical wind direction real-time data y of a target unit at the jth sampling moment before the current prediction momentt-jHistorical wind direction real-time data z of the 1 st associated unit at the jth sampling moment before the current prediction momentt-jThe historical wind direction real-time data of the mth associated unit at the jth sampling time before the current prediction time is obtained, j is 1, 2, … … p, and m is the total number of the associated units;
the vector autoregressive prediction model is then:
wherein x istFor the predicted value of the wind direction of the target unit at the current prediction time, w0Is a constant value, w0Representing the stationary terms of the historical wind direction time series of the target unit, αjβ co-order co-integration influence factor of the historical wind direction real-time data of the j sampling moment before the current prediction moment of the target unit per se on the wind direction prediction value of the target unit at the current prediction momentjRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the 1 st associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment, and so on, gammajRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the mth associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment,tis a random perturbation term with a mean of 0 and a constant variance, αj、βj、γjAndtare all constants obtained by pre-training.
5. The wind direction information prediction method according to any one of claims 1 to 4, wherein the method is used for yaw control of the target unit after acquiring the wind direction prediction value of the target unit at the current prediction time.
6. A wind direction information prediction system, comprising:
data acquisition unit (1): the unit is used for acquiring historical wind direction real-time data of each wind turbine in the wind power plant;
associated unit screening unit (2): the unit screens a plurality of associated units of a target unit according to historical wind direction real-time data;
data extraction unit (3): the unit extracts historical wind direction real-time data of a target unit and an associated unit at p sampling moments before the current prediction moment to form a historical wind direction time sequence of the target unit and a historical wind direction time sequence of the associated unit, wherein p is a constant;
prediction unit (4): the unit inputs the historical wind direction time sequence of the target unit and the historical wind direction time sequence of the associated unit into a pre-established prediction model to obtain a wind direction prediction value of the target unit at the current prediction time.
7. A wind direction information prediction system according to claim 6, characterized in that said associated crew filtering unit (2) comprises:
pretreatment subunit (21): the sub-unit performs the following operations,
firstly, a recycling process:
let the original dataset be Xr＝{x1,x2,…,xnObtaining a recycling processing data set X 'after recycling processing'r＝{x′1,x′2,…,x′nAnd (c) the step of (c) in which,
x′1＝x1，
when the value of i is greater than 1,
wherein x isiFor the ith historical wind direction real-time data, xi' is the ith historical wind direction real-time data after the circulation processing,i is 1, 2, … … n, n is the total number of data in the original data set;
then, resampling:
setting a resampling interval as T seconds, wherein each resampling interval corresponds to d data in the data set for de-circulation processing, and sequentially dividing each d data in the data set for de-circulation processing into a group and calculating an average value to obtain the real-time data of the re-sampled historical wind direction;
screening subunit (22): the subunit screens the associated units based on the resampled historical wind direction real-time data.
8. A wind direction information prediction system according to claim 6, characterized in that the prediction model in the prediction unit (4) is a vector autoregressive prediction model.
9. The system of claim 8, wherein the vector autoregressive prediction model is mathematically expressed as:
setting a historical wind direction time sequence of the target unit as follows: x ═ Xt-p,xt-p+1,…,xt-2,xt-1}，
The 1 st associated unit historical wind direction time sequence is as follows: y ═ Yt-p,yt-p+1,…,yt-2,yt-1}，
By analogy, the historical wind direction time sequence of the mth associated unit is as follows: z ═ Zt-p,zt-p+1,…,zt-2,zt-1}，
Wherein x ist-jHistorical wind direction real-time data y of a target unit at the jth sampling moment before the current prediction momentt-jHistorical wind direction real-time data z of the 1 st associated unit at the jth sampling moment before the current prediction momentt-jThe historical wind direction real-time data of the mth associated unit at the jth sampling time before the current prediction time is obtained, j is 1, 2, … … p, and m is the total number of the associated units;
the vector autoregressive prediction model is then:
wherein x istFor the predicted value of the wind direction of the target unit at the current prediction time, w0Is a constant value, w0Representing the stationary terms of the historical wind direction time series of the target unit, αjβ co-order co-integration influence factor of the historical wind direction real-time data of the j sampling moment before the current prediction moment of the target unit per se on the wind direction prediction value of the target unit at the current prediction momentjRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the 1 st associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment, and so on, gammajRepresenting the co-order co-integration influence factor of the historical wind direction real-time data of the mth associated unit at the jth sampling moment before the current prediction moment on the wind direction prediction value of the target unit at the current prediction moment,tis a random perturbation term with a mean of 0 and a constant variance, αj、βj、γjAndtare all constants obtained by pre-training.
10. The wind direction information prediction system according to claims 6 to 9, wherein the system is connected to a fan yaw control system, and the prediction unit (4) obtains the wind direction predicted value of the target unit at the current prediction time and then uses the wind direction predicted value for yaw control of the target unit in the fan yaw control system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010397191.5A CN111680823A (en) | 2020-05-12 | 2020-05-12 | Wind direction information prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010397191.5A CN111680823A (en) | 2020-05-12 | 2020-05-12 | Wind direction information prediction method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111680823A true CN111680823A (en) | 2020-09-18 |
Family
ID=72433543
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010397191.5A Pending CN111680823A (en) | 2020-05-12 | 2020-05-12 | Wind direction information prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111680823A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112798949A (en) * | 2020-10-22 | 2021-05-14 | 国家电网有限公司 | Pumped storage unit generator temperature early warning method and system |
| CN113374634A (en) * | 2021-07-01 | 2021-09-10 | 浙江浙能技术研究院有限公司 | Wind turbine yaw wind alignment method under anemoscope fault mode |
| CN113850418A (en) * | 2021-09-02 | 2021-12-28 | 支付宝(杭州)信息技术有限公司 | Method and device for detecting abnormal data in time sequence |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107408144A (en) * | 2014-11-14 | 2017-11-28 | Zoll医疗公司 | Medical precursor event estimation |
| CN108547736A (en) * | 2018-03-27 | 2018-09-18 | 中南大学 | The Yaw control method of wind speed and direction prediction technique and wind power generating set |
| CN110296055A (en) * | 2019-06-10 | 2019-10-01 | 同济大学 | A kind of wind direction prediction association seed unit screening technique |
| CN110598911A (en) * | 2019-08-21 | 2019-12-20 | 同济大学 | Wind speed prediction method for wind turbine of wind power plant |
-
2020
- 2020-05-12 CN CN202010397191.5A patent/CN111680823A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107408144A (en) * | 2014-11-14 | 2017-11-28 | Zoll医疗公司 | Medical precursor event estimation |
| CN108547736A (en) * | 2018-03-27 | 2018-09-18 | 中南大学 | The Yaw control method of wind speed and direction prediction technique and wind power generating set |
| CN110296055A (en) * | 2019-06-10 | 2019-10-01 | 同济大学 | A kind of wind direction prediction association seed unit screening technique |
| CN110598911A (en) * | 2019-08-21 | 2019-12-20 | 同济大学 | Wind speed prediction method for wind turbine of wind power plant |
Non-Patent Citations (4)
| Title |
|---|
| （英）布赖恩•斯诺登，霍华德•R.文编；安佳等译: "《宏观经济学百科词典》", 31 January 2019, 江苏人民出版社 * |
| 孙春顺: "小时风速的向量自回归模型及应用", 《中国电机工程学报》 * |
| 宝乐尔其木格: "中国沿海风特性研究", 《中国优秀博硕士学位论文全文数据库(硕士)基础科学辑》 * |
| 杨茂，杨春霖: "考虑风向的风电功率预测不确定性研究", 《东北电力大学学报》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112798949A (en) * | 2020-10-22 | 2021-05-14 | 国家电网有限公司 | Pumped storage unit generator temperature early warning method and system |
| CN113374634A (en) * | 2021-07-01 | 2021-09-10 | 浙江浙能技术研究院有限公司 | Wind turbine yaw wind alignment method under anemoscope fault mode |
| CN113374634B (en) * | 2021-07-01 | 2022-09-06 | 浙江浙能技术研究院有限公司 | Wind turbine yaw wind alignment method under anemoscope fault mode |
| CN113850418A (en) * | 2021-09-02 | 2021-12-28 | 支付宝(杭州)信息技术有限公司 | Method and device for detecting abnormal data in time sequence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111680823A (en) | Wind direction information prediction method and system | |
| CN107194625B (en) | Wind power plant wind curtailment electric quantity evaluation method based on neural network | |
| CN112348292B (en) | Short-term wind power prediction method and system based on deep learning network | |
| CN109992921A (en) | A kind of online soft sensor method and system of the coal-fired plant boiler thermal efficiency | |
| CN110417011A (en) | A kind of online dynamic secure estimation method based on mutual information Yu iteration random forest | |
| CN106709641A (en) | Monte-Carlo simulation based small interference probability risk analysis and simulation method | |
| CN109274121B (en) | Wind power plant control parameter optimization method and system | |
| CN110991737A (en) | Ultra-short-term wind power prediction method based on deep belief network | |
| CN105550747A (en) | Sample training method for novel convolutional neural network | |
| CN109063930A (en) | A kind of dynamic wind power plant general power prediction technique based on clustering | |
| CN116306306A (en) | Micro-grid inertia constant estimation method based on African bald-Condition algorithm | |
| CN108052963A (en) | The data screening method, apparatus and wind power generating set of wind power prediction modeling | |
| CN110096730A (en) | A kind of network voltage fast evaluation method and system | |
| CN110943473A (en) | Generator coherence identification method based on wide area measurement system and clustering theory | |
| Liu et al. | A dynamic clustering model of wind farm based on the operation data | |
| CN101929684A (en) | Coal consumption calculating method of composite firing low heat value gas unit | |
| CN116167740A (en) | Digital twin model construction method for intelligent maintenance of pumped storage power station | |
| CN106326982B (en) | A kind of more meteorological wind speed fusion methods optimized based on neural network and pattern search | |
| CN107102543A (en) | The forming method and device of a kind of energy router anti-interference controller | |
| CN113962357A (en) | GWO-WNN-based distributed photovoltaic power data virtual acquisition method | |
| CN113054653A (en) | Power system transient stability evaluation method based on VGGNet-SVM | |
| CN110222361A (en) | A kind of Power Output for Wind Power Field analogy method and device | |
| Ma et al. | GAN based data analysis and mining for smart shop floor scheduling | |
| CN107124003A (en) | Wind power plant wind energy Forecasting Methodology and equipment | |
| CN112688330B (en) | Method and system for generating power grid analysis state library based on D5000 online power flow |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200918 |