CN109102101B - Wind speed prediction method and system for wind power plant - Google Patents
Wind speed prediction method and system for wind power plant Download PDFInfo
- Publication number
- CN109102101B CN109102101B CN201710474284.1A CN201710474284A CN109102101B CN 109102101 B CN109102101 B CN 109102101B CN 201710474284 A CN201710474284 A CN 201710474284A CN 109102101 B CN109102101 B CN 109102101B
- Authority
- CN
- China
- Prior art keywords
- wind
- wind speed
- data
- linear regression
- program module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000012417 linear regression Methods 0.000 claims abstract description 57
- 230000009467 reduction Effects 0.000 claims abstract description 45
- 238000010219 correlation analysis Methods 0.000 claims abstract description 29
- 230000006870 function Effects 0.000 claims description 51
- 238000012545 processing Methods 0.000 claims description 34
- 238000012549 training Methods 0.000 claims description 33
- 238000012360 testing method Methods 0.000 claims description 27
- 238000004458 analytical method Methods 0.000 claims description 13
- 230000002787 reinforcement Effects 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000011478 gradient descent method Methods 0.000 claims description 4
- 230000008901 benefit Effects 0.000 abstract description 3
- 230000000875 corresponding effect Effects 0.000 description 18
- 238000004422 calculation algorithm Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 9
- 239000011159 matrix material Substances 0.000 description 7
- 239000013598 vector Substances 0.000 description 6
- 238000000513 principal component analysis Methods 0.000 description 4
- 238000000611 regression analysis Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 238000009795 derivation Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 229910000831 Steel Inorganic materials 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 125000004432 carbon atom Chemical group C* 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000008014 freezing Effects 0.000 description 1
- 238000007710 freezing Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000010959 steel Substances 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
- 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
-
- H02J3/386—
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- 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
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A30/00—Adapting or protecting infrastructure or their operation
-
- 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
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/70—Wind energy
- Y02E10/76—Power conversion electric or electronic aspects
-
- 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
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Wind Motors (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method and a system for predicting wind speed of a wind power plant, wherein the method for predicting the wind speed of the wind power plant comprises the following steps: acquiring historical data of a wind power plant; performing dimensionality reduction on the acquired historical data; carrying out wind speed correlation analysis on the data subjected to the dimensionality reduction treatment; and establishing a linear regression model according to the wind speed correlation analysis result, and predicting the wind speed of the specific wind generating set in the wind power plant based on the established linear regression model. The method can accurately predict the wind speed of a certain specific wind generating set, so that the wind speed and the wind direction instrument can still keep high wind power efficiency after the fault of the wind speed and the wind direction instrument, can effectively reduce the downtime as an emergency system, improve the availability of the wind generating set, increase the unit output of the wind generating set, and has obvious economic benefit.
Description
Technical Field
The invention relates to the technical field of wind power, in particular to a method and a system for predicting wind speed of a wind power plant.
Background
Wind power technology has now become a major contribution to the world's growing clean power market, and accurate and reliable wind power predictions are widely recognized as a major cause of increasing wind power generation penetration rates. Because aerogenerator moves in outdoor high altitude, and anemorumbometer is outside the unit cabin cover, its operational environment is very complicated: not only in the face of various extreme conditions such as vibration, dust, insolation, freezing, rain and the like, but also salt fog corrosion of offshore and mudflat wind fields is suffered, and the probability of failure of the anemorumbometer is relatively high.
Generally, the emergency control method for wind speed and wind direction failure of a wind driven generator in the prior art is as follows: all units in the wind farm take the wind farm as a center, and form a small network with three units closest to other geographical positions through a data acquisition and supervisory control and data acquisition (SCADA) system of the wind farm; when the unit normally operates, the unit feeds back the collected wind speed and direction signals to the central controller through the SCADA system, and the central controller displays various information of the unit including wind speed and direction in real time; when the anemorumbometer of the unit fails, firstly, the anemorumbometer reports to a central processing unit, the central processing unit displays the different colors at the remarkable positions and gives a warning to monitoring personnel; meanwhile, the system assumes that the wind speed and the wind direction of 3 wind generating sets are similar, the data of the wind speed and the wind direction instrument of the 3 wind generating sets are returned to the central controller, after the central controller confirms that the information of 1-2 wind generating sets is valid after screening, the SCADA system selects the wind speed and the wind direction data according to the sequence of the first set, the second set and the third set with the priority and sends the data to the central set, and therefore the wind generating sets can obtain the wind speed and the wind direction data to continuously operate after the wind speed and the wind direction instrument are in failure.
However, in the prior art, the wind speed and the wind direction of adjacent wind generating sets are only assumed to be the same, emergency control is performed on the failure of the anemorumbometer of the wind generating set, the wind speed change or attenuation after a certain distance is not considered, the prediction deviation is large, and the safe operation of the wind generating set is difficult to meet.
Disclosure of Invention
The invention provides a method and a system for predicting wind speed of a wind power plant, which can more accurately predict the wind speed of a certain specific wind generating set by performing dimensionality reduction processing and correlation analysis on historical data of the wind power plant and performing regression analysis modeling based on a machine learning method.
One aspect of the invention provides a method for predicting wind speed of a wind farm, comprising the steps of: acquiring historical data of a wind power plant; performing dimensionality reduction on the acquired historical data; carrying out wind speed correlation analysis on the data subjected to the dimensionality reduction treatment; and establishing a linear regression model according to the wind speed correlation analysis result, and predicting the wind speed of the specific wind generating set in the wind power plant based on the established linear regression model.
Preferably, the historical data comprises historical wind speed data and historical operating condition data.
Preferably, the step of performing dimension reduction processing on the acquired historical data includes: centralizing the acquired historical data and solving a target function according to a centralization processing result; forming a maximum solving problem by the objective function and the constraint condition; constructing a Lagrange function, and solving a maximization solution problem to obtain a maximum characteristic value; and obtaining a result of dimension reduction processing according to the maximum eigenvector corresponding to the maximum eigenvalue.
Preferably, the step of performing wind speed correlation analysis on the data after the dimensionality reduction processing comprises: preprocessing the data after the dimensionality reduction; dividing the preprocessed data into training samples and test tuples; respectively calculating Euler distances of the test tuples and the training tuples, and taking K training tuples with the minimum distance; setting the label of the K training tuples which appears most times as the category of the test tuple; dividing the training sample into at least one region according to the type of the test tuple; and calculating an error rate, and selecting a K value corresponding to the minimum error rate, wherein the K value corresponding to the minimum error rate indicates that the wind speeds of K wind generating sets in the wind power plant have linear correlation.
Preferably, the step of establishing a linear regression model according to the wind speed correlation analysis result and predicting the wind speed of a specific wind generating set in the wind farm based on the established linear regression model comprises the following steps: respectively carrying out linear regression analysis on the wind speed data in different areas; establishing a linear regression model according to the linear regression analysis result, and establishing a loss function through the model and historical data; and solving the minimum value of the loss function, establishing a linear regression model according to the parameters corresponding to the minimum value of the loss function, and predicting the wind speed of the specific wind generating set in the wind power plant based on the established linear regression model.
Preferably, the minimum value of the loss function is obtained by a least square method and/or a gradient descent method.
Preferably, the method further comprises determining whether the predicted wind speed is biased according to a wind power curve of the specific wind turbine generator set.
Preferably, the method further comprises the step of performing reinforcement learning on the established linear regression model.
Another aspect of the invention provides a system for predicting wind speed of a wind farm, the system comprising: the data processing program module is used for acquiring historical data of the wind power plant and performing dimension reduction processing on the acquired historical data; the correlation analysis program module is used for carrying out wind speed correlation analysis on the data subjected to the dimensionality reduction treatment; and the wind speed prediction program module is used for establishing a linear regression model according to the wind speed correlation analysis result and predicting the wind speed of a specific wind generating set in the wind power plant based on the established linear regression model.
Preferably, the historical data in the data processing program module comprises historical wind speed data and historical operating condition data.
Preferably, the data processing program module includes: the target function program module is used for carrying out centralized processing on the acquired historical data and solving a target function according to a centralized processing result; the maximum characteristic value program module is used for forming a maximum solving problem by the target function and the constraint condition together, constructing a Lagrangian function and solving the maximum solving problem to obtain a maximum characteristic value; and the dimension reduction processing program module is used for obtaining a dimension reduction processing result according to the maximum eigenvector corresponding to the maximum eigenvalue.
Preferably, the correlation analysis program module includes: the data preprocessing program module is used for preprocessing the data subjected to the dimensionality reduction processing; the Euler distance program module divides the preprocessed data into a training sample and a testing tuple and respectively calculates the Euler distance between the testing tuple and the training tuple; the region dividing program module is used for taking K training tuples with the minimum Euler distance, setting a label which appears most times in the K training tuples as a type of a test tuple, and dividing a training sample into at least one region according to the type of the test tuple; and the linear correlation program module is used for calculating the error rate and selecting a K value corresponding to the minimum error rate, wherein the K value corresponding to the minimum error rate indicates that the wind speeds of K wind generating sets in the wind power plant have linear correlation.
Preferably, the wind speed prediction program module comprises: the model building program module is used for respectively carrying out linear regression analysis on the wind speed data in different areas and building a linear regression model according to the linear regression analysis result; and the wind speed prediction program module is used for constructing a loss function according to the linear regression model and the historical data, solving parameters corresponding to the minimum value of the loss function, and predicting the wind speed of a specific wind generating set in the wind power plant based on the solved parameters and the established linear regression model.
Preferably, the minimum value of the loss function is obtained by a least square method and/or a gradient descent method.
Preferably, the method further comprises the following steps: and the deviation testing program module is used for testing whether the predicted wind speed has deviation or not according to the wind power curve of the specific wind generating set.
Preferably, the method further comprises the following steps: and the reinforcement learning program module is used for reinforcement learning of the established linear regression model.
Another aspect of the invention provides a computer-readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform a method of predicting wind speed for a wind farm as described above.
Another aspect of the invention provides a computer device comprising a processor and a memory storing a computer program which, when executed by the processor, performs a method of predicting wind speed for a wind farm as described above.
According to the invention, through processing and analyzing historical data, and using correlation analysis and regression analysis modeling based on machine learning, the wind speed of a certain specific wind generating set is accurately predicted, so that the wind speed and the wind direction instrument still keep high generating efficiency after a fault, the downtime is effectively reduced, and the availability of the wind generating set is improved.
Drawings
The above and other aspects, features and advantages of exemplary embodiments of the present invention will become more apparent from the following description taken in conjunction with the accompanying drawings, in which:
FIG. 1 shows a flow chart of a method of predicting wind speed of a wind farm according to an embodiment of the invention;
FIG. 2 illustrates the results of a correlation analysis using a nearest neighbor node (KNN) algorithm, according to an embodiment of the present invention;
FIG. 3 illustrates wind speed prediction results for a particular wind generating set in a wind farm according to an embodiment of the present invention;
FIG. 4 shows a block diagram of a prediction system for wind speed of a wind farm according to an embodiment of the invention;
FIG. 5 shows a block diagram of data handler modules, according to an embodiment of the invention;
FIG. 6 illustrates a block diagram of a relevance analysis program module according to an embodiment of the invention;
FIG. 7 shows a block diagram of wind speed prediction program modules according to an embodiment of the invention.
In the drawings, like reference numerals will be understood to refer to like elements, features and structures.
Detailed Description
The following description with reference to the accompanying drawings is provided to assist in a comprehensive understanding of exemplary embodiments of the invention as defined by the claims and their equivalents. The following description with reference to the figures includes various specific details to aid understanding, but the specific details are to be considered exemplary only. Accordingly, those of ordinary skill in the art will appreciate that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the invention. In addition, descriptions of well-known functions and constructions may be omitted for clarity and conciseness.
FIG. 1 is a flow chart illustrating a method of predicting wind speed for a wind farm according to an embodiment of the present invention.
As shown in fig. 1, first, in step S100, historical data of a wind farm is acquired. Specifically, obtaining historical data for the wind farm may include historical wind speed data and historical operating condition data. According to an embodiment of the invention, historical wind speed data and historical operating condition data of a wind farm may be read from a SCADA system, for example. The historical working condition data comprises data which affect the operation of the wind generating set of the wind power plant, such as the yaw angle, the pitch angle and the temperature of the wind generating set, inconsistent and wrong data items in the historical data set are removed, and normalization processing is carried out on the data.
Next, in step S200, the acquired history data is subjected to dimension reduction processing. Specifically, the historical wind speed data and the historical working condition data are subjected to dimensionality reduction. According to the embodiment of the invention, through carrying out dimension reduction processing on the historical data, the historical operating data of a large amount of wind generating sets is simplified under the condition of little loss of data characteristics, so that a wind speed prediction model can be established more quickly and accurately. According to an embodiment of the invention, a Principal Component Analysis (PCA) algorithm may be used for data dimensionality reduction.
In wind speed prediction for a wind park, the data is denoted as vectors, and the wind speed and other relevant information for a particular wind park at a certain time of day form a set, where the time is a recorded flag and not a measure. According to the embodiment of the invention, the wind speed is required to be known about which measurement values are related, and due to the numerous working conditions of the wind generating set, the data are subjected to dimensionality reduction. The dimension reduction processing can cause data loss, but the actual data has correlation, so that the loss of information can be reduced as much as possible while the dimension is reduced. For example, a given data set is
The centralization treatment is carried out on the steel plate, and the centralization result is as follows:
in the above formula, the first and second carbon atoms are,
representing fan operating data including wind speed, wind direction, etc.,
represents each oneThe average of the column data is,
represent
And the average value
The difference of (a). The most spread of the centralized data in the direction of the first principal axis u1, that is, the sum of the absolute values of the projections in the direction of u1 is the largest, the projection is calculated by taking the inner product of x and u1, and u1 is a unit vector, and the objective function is obtained as
Converting the objective function to obtain
And with constraint conditions
And constructing a maximum solution problem.
Constructing lagrange functions
For u is paired1Derivation:
XXTu1＝λu1
the derivation result of the above equation is substituted into the objective function to obtain:
the eigenvector corresponding to the eigenvalue lambda is known as u1And obtaining the maximum characteristic value and obtaining the corresponding maximum characteristic vector according to the maximum characteristic value. Wherein the feature vector u1And the wind speed, the wind direction and other fan operation data are shown. According to an embodiment of the present invention, historical data acquired at 2016, 1, month and time is selectedThe row matrix is 11572 × 270 matrix, the first 5 maximum eigenvalues are selected from the obtained eigenvalues, the corresponding maximum eigenvector is obtained according to the maximum eigenvalue, the maximum eigenvector forms a 270 × 5 eigenvector matrix, finally, the sample data is multiplied by the eigenvector matrix to obtain 11572 × 5 matrix, namely, the wind speed data and the working condition data with correlation in the historical data are obtained through PCA dimension reduction processing, wherein the historical working condition data obtained after dimension reduction are few, and the historical working condition with few data is ignored in the next calculation. The historical data of the wind generating set achieves the result of dimension reduction and simplification under the condition of extremely little loss of data characteristics.
In step S300, wind speed correlation analysis is performed on the data after the dimensionality reduction processing. Specifically, the wind speed data after the dimensionality reduction processing in step S200 is preprocessed, and the preprocessed data is subjected to wind speed correlation analysis. According to the embodiment of the invention, a nearest neighbor node (KNN) algorithm is used for correlation analysis.
Due to the large wind field area in general, the wind speeds of all wind turbine generators are not correlated, and only a few wind turbine generators may have correlation in wind speed due to topographic reasons. If the difficulty of calculating the aerohydrodynamics is high by establishing a physical model and the calculation can be performed only one by one for different wind fields, the KNN algorithm is selected and used in the invention, and the mathematical model is constructed to deduce which wind speed of the wind generating sets is relatively close.
Specifically, the data after dimensionality reduction by the PCA algorithm is divided into a training sample and a testing tuple, and the distances between the testing tuple and the training tuple are respectively calculated based on the Euclidean theorem, wherein the Euclidean distance formula is as follows:
in the above formula, xjRepresenting test tuples, yjRepresenting a training tuple. Then, according to K training tuples with the minimum distance in the distances calculated by the formula, judging the K training tuples with the maximum occurrence frequencyAnd setting the labels as the types of the test tuples, and dividing the training sample into at least one region according to the types of the test tuples. And finally, calculating an error rate, and selecting a K value corresponding to the minimum error rate, wherein the K value corresponding to the minimum error rate indicates that the wind speeds of K wind generating sets in the wind power plant have linear correlation. According to the embodiment of the invention, 16 wind generating sets are taken as training samples, are respectively numbered, and are subjected to correlation analysis through a nearest node (KNN) algorithm. The result of the correlation analysis using the nearest neighbor node (KNN) algorithm according to an embodiment of the present invention will be described in detail with reference to fig. 2.
Fig. 2 is a diagram illustrating the results of a correlation analysis using a nearest neighbor node (KNN) algorithm according to an embodiment of the present invention.
As can be seen from fig. 2, the training sample is divided into four regions after correlation analysis by a nearest neighbor node (KNN) algorithm, wherein wind turbine generators with correlation numbers 01, 02, 03, 04, 05, 08 are classified as the same region, wind turbine generators with correlation numbers 06 and 07 are classified as the same region, wind turbine generators with correlation numbers 10, 11, 12, 13 are classified as the same region, and wind turbine generators with correlation numbers 09, 14, 15, 16 are classified as the same region. Returning to fig. 1, in step S400, a linear regression model is established according to the wind speed correlation analysis result, and the wind speed of a specific wind turbine generator set in the wind farm is predicted based on the established linear regression model. Specifically, linear regression analysis is performed on different areas divided by the wind speed correlation analysis in the step S300, a linear regression model is established, and finally, the wind speed of the wind farm is predicted based on the linear regression model. According to the embodiment of the present invention, linear regression analysis is performed on the wind speed data in the four different regions obtained in fig. 2, for example, the wind speed data in the regions consisting of the wind turbine generators numbered 01, 02, 03, 04, 05, and 08. In the acquired historical data, each component is regarded as feature data, each feature at least corresponds to one unknown parameter, a linear regression model function is constructed according to the regression analysis result, and the vector expression form is as follows:
hθ(x)＝θTX
wherein theta is an unknown parameter, X is historical wind speed data and historical working condition data after dimensionality reduction, and hθ(x) Representing the predicted wind speed of the wind farm. And then constructing a loss function according to the linear regression model function and the historical data to evaluate the h function, wherein the loss function is expressed as follows:
wherein, y(i)Representing the actual value h of the historical wind speed data after dimensionality reductionθ(x(i)) The predicted wind speed value through the model is shown, and the value of theta is adjusted so that the loss function J (theta) takes the minimum value, wherein the minimum value of the loss function is obtained through the least square method and/or gradient descent. For example, when the least square method is selected to solve the minimum value of the loss function, the historical wind speed data and the historical working condition data after dimensionality reduction are represented as an X matrix, the predicted wind speed result of the wind power plant is represented as a y vector, and X is a column full-rank matrix to obtain parameters
And finally, respectively carrying out parameter theta assignment on the historical wind speed data and the historical working condition data according to the optimal parameter theta corresponding to the minimum value of the loss function and the factor influencing the wind speed of the wind power plant, and predicting the wind speed of the wind power plant through a linear regression model function.
In step S500, it is determined whether the predicted wind speed deviates according to the wind power curve of the specific wind turbine. Specifically, according to the embodiment of the invention, for example, when a certain wind generating set in a wind farm has a fault, the prediction method of the invention calculates the wind speed of the wind generating set with a failed anemorumbometer according to the correlation model and the wind speed model of the wind farm, and simultaneously the central control system monitors the wind power curve of the wind generating set, if the wind power curve is abnormal, the wind speed is considered to have a large deviation, the wind generating set is immediately controlled to stop to prevent a safety accident, if the wind power curve is not abnormal, the central control system is considered to determine that the wind speed prediction system is accurate in calculation, and the wind generating set is kept running under the condition that the anemorumbometer has the fault.
FIG. 3 is a graph illustrating wind speed predictions for a particular wind generating set in a wind farm according to an embodiment of the present invention.
As shown in fig. 3, linear regression analysis is performed on wind speed data of wind generating sets in areas to which the wind generating sets with numbers 01, 02, 03, 04, 05, and 08 belong, wherein the abscissa is time t, and the ordinate represents speed v/s, and assuming that the anemorumbometer of the wind generating set No. 01 fails, the predicted wind speed of the wind generating set No. 01 is obtained by applying the model of the wind speed prediction method for the wind farm provided by the invention, and as can be seen from fig. 3, the error between the obtained predicted wind speed and the actual wind speed is within 10%, and the error is relatively small.
According to the embodiment of the invention, the invention further comprises the step of performing reinforcement learning on the established linear regression model. Specifically, under the condition that the anemorumbometer of the wind generating set normally operates, wind speed data are continuously collected and accumulated, wind speed is continuously predicted according to the prediction method of the invention, a predicted value is compared with an actual value, and a loss function value of a fitting curve is continuously reduced along with the continuous increase of the data quantity, so that the prediction method continuously enhances algorithm learning in the operation process to be close to the actual value as much as possible.
FIG. 4 is a block diagram illustrating a prediction system for wind speed for a wind farm according to an embodiment of the present invention.
As shown in FIG. 4, wind farm wind speed prediction system 600 may include a data processing program module 601, a correlation analysis program module 602, and a wind speed prediction program module 603. According to embodiments of the invention, the system 600 for predicting wind speed of a wind farm may be implemented by various computing devices (e.g., computers, servers, workstations, etc.). Specifically, the data processing program module 601 is configured to obtain historical data of the wind farm and perform dimension reduction processing on the obtained historical data. And the correlation analysis program module 602 is used for performing wind speed correlation analysis on the data after the dimensionality reduction processing. The wind speed prediction program module 603 is configured to build a linear regression model according to the wind speed correlation analysis result and predict the wind speed of a specific wind turbine generator set in the wind farm based on the built linear regression model.
The data handler module 601 according to an embodiment of the present invention will be described in detail with reference to fig. 5.
FIG. 5 shows a block diagram of data processing program modules according to an embodiment of the invention.
As shown in fig. 5, the data handler module 601 includes an objective function program module 101, a maximum feature value program module 102, and a dimension reduction handler module 103. The target function program module 101 performs centralized processing on the acquired historical data and obtains a target function according to a centralized processing result, the maximum characteristic value program module 102 forms a maximum solving problem together with the target function obtained by the target function program module 101 and constraint conditions, constructs a Lagrangian function, and solves the maximum solving problem to obtain a maximum characteristic value; the dimension reduction processing program module 103 obtains a dimension reduction processing result according to the maximum eigenvector corresponding to the maximum eigenvalue.
The correlation analysis program module 602 according to an embodiment of the present invention will be described in detail with reference to fig. 6.
FIG. 6 illustrates a block diagram of a relevance analysis program module according to an embodiment of the invention.
As shown in FIG. 6, the correlation analysis program module 602 includes a data preprocessing program module 104, a Euler distance program module 105, a region partitioning program module 16, and a linear correlation program module 107. The data preprocessing program module 104 is used for preprocessing data after dimensionality reduction, the euler distance program module 105 is used for dividing the preprocessed data into training samples and testing tuples and respectively calculating euler distances between the testing tuples and the training tuples, the region division program module 106 is used for taking K training tuples with the minimum distance according to the euler distances calculated by the euler distance program module 105, setting labels which appear most times in the K training tuples as the classes of the testing tuples, dividing the training samples into at least one region according to the classes of the testing tuples, the linear correlation program module 107 is used for calculating an error rate and selecting a K value corresponding to the minimum error rate, wherein the K value corresponding to the minimum error rate indicates that the wind speeds of K wind generating sets in the wind power plant have linear correlation.
The wind speed prediction program module 603 according to an embodiment of the invention will be described in detail below with reference to fig. 7.
FIG. 7 shows a block diagram of wind speed prediction program modules according to an embodiment of the invention.
As shown in FIG. 7, wind speed prediction program module 603 includes a model building program module 108 and a wind speed prediction program module 109. The model building program module 108 performs linear regression analysis on the wind speed data in different regions respectively, a linear regression model is built according to the linear regression analysis result, the wind speed prediction program module 109 builds a loss function according to the linear regression model built by the model building program module 108 and historical data, parameters corresponding to the minimum value of the loss function are solved, and the wind speed of a specific wind generating set in the wind farm is predicted based on the solved parameters and the built linear regression model. Wherein, the minimum value of the loss function is obtained by a least square method and/or a gradient descent method.
According to an embodiment of the invention, the system 600 for predicting wind speed of a wind farm further comprises: and a deviation test program module 604 for determining whether the predicted wind speed is deviated according to the wind power curve of the specific wind generating set.
According to an embodiment of the invention, the system 600 for predicting wind speed of a wind farm further comprises: and the reinforcement learning program module 605 is configured to continuously perform reinforcement learning on the established linear regression model, so that the prediction system continuously reinforces algorithm learning to approach the actual value as much as possible during the operation process.
According to the method and the system for predicting the wind speed of the wind power plant, the wind speed of a certain specific wind generating set can be accurately predicted by analyzing historical data and modeling the wind speed of the wind power plant and modeling the wind speed by using correlation analysis and regression analysis based on machine learning, so that the wind speed of the certain specific wind generating set can still keep high wind power efficiency after a anemorumbometer fails, the method and the system can be used as an emergency system, the downtime can be effectively reduced, the availability of the wind generating set is improved, the unit output of the wind generating set is increased, and the economic benefit is obvious.
A method of predicting wind speed of a wind farm according to embodiments of the present invention may be implemented as computer readable code on a computer readable recording medium or may be transmitted through a transmission medium. The computer readable recording medium is any data storage device that can store data which can be thereafter read by a computer system. The computer-readable storage medium stores a computer program which, when executed by a processor, performs the method of predicting wind speed of a wind farm shown in fig. 1. Examples of the computer readable recording medium include, but are not limited to, read-only memory (ROM), random-access memory (RAM), Compact Disc (CD) -ROM, Digital Versatile Disc (DVD), magnetic tape, floppy disk, optical data storage device. The transmission medium may include a carrier wave transmitted over a network or various types of communication channels. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.
Another embodiment of the invention provides a computer arrangement comprising a processor and a memory storing a computer program which, when executed by the processor, performs the method of predicting wind speed for a wind farm as shown in fig. 1.
While the invention has been shown and described with reference to certain exemplary embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims and their equivalents.
Claims (12)
1. A method for predicting wind speed of a wind power plant is characterized by comprising the following steps:
acquiring historical data of a wind power plant;
performing dimensionality reduction on the acquired historical data;
carrying out wind speed correlation analysis on the data subjected to the dimensionality reduction treatment;
establishing a linear regression model according to the wind speed correlation analysis result, predicting the wind speed of a specific wind generating set in the wind power plant based on the established linear regression model,
wherein, the step of performing wind speed correlation analysis on the data after the dimensionality reduction processing comprises the following steps:
preprocessing the data after the dimensionality reduction;
dividing the preprocessed data into a training sample and a testing tuple, and respectively calculating Euler distances between the testing tuple and the training tuple;
taking K training tuples with minimum Euler distances, setting labels which appear most times in the K training tuples as the types of the testing tuples, and dividing the training sample into at least one region according to the types of the testing tuples;
calculating an error rate, and selecting a K value corresponding to the minimum error rate, wherein the K value corresponding to the minimum error rate indicates that the wind speeds of K wind generating sets in the wind power plant have linear correlation,
the steps of establishing a linear regression model according to the wind speed correlation analysis result and predicting the wind speed of a specific wind generating set in the wind power plant based on the established linear regression model comprise:
respectively carrying out linear regression analysis on the wind speed data in different areas, and establishing a linear regression model according to the linear regression analysis result;
constructing a loss function according to the linear regression model and the historical data, solving parameters corresponding to the minimum value of the loss function, predicting the wind speed of a specific wind generating set in the wind power plant based on the solved parameters and the established linear regression model,
wherein the historical data comprises historical wind speed data and historical operating condition data.
2. The prediction method of claim 1, wherein the step of performing dimension reduction processing on the acquired historical data comprises:
centralizing the acquired historical data and solving a target function according to a centralization processing result;
the objective function and the constraint condition jointly form a maximum solving problem, a Lagrangian function is constructed, and the maximum solving problem is solved to obtain a maximum characteristic value;
and obtaining a result of dimension reduction processing according to the maximum eigenvector corresponding to the maximum eigenvalue.
3. The prediction method of claim 1, wherein the loss function minimum is found by a least squares method and/or a gradient descent method.
4. The prediction method of claim 1, further comprising:
and determining whether the predicted wind speed is deviated or not according to the wind power curve of the specific wind generating set.
5. The prediction method of claim 1, further comprising: and performing reinforcement learning on the established linear regression model.
6. A prediction system for wind speed in a wind farm, the prediction system comprising:
the data processing program module is used for acquiring historical data of the wind power plant and performing dimension reduction processing on the acquired historical data;
the correlation analysis program module is used for carrying out wind speed correlation analysis on the data subjected to the dimensionality reduction treatment;
a wind speed prediction program module for establishing a linear regression model according to the wind speed correlation analysis result and predicting the wind speed of a specific wind generating set in the wind farm based on the established linear regression model,
wherein the relevance analysis program module comprises:
the data preprocessing program module is used for preprocessing the data subjected to the dimensionality reduction processing;
the Euler distance program module divides the preprocessed data into a training sample and a testing tuple and respectively calculates the Euler distance between the testing tuple and the training tuple;
the region dividing program module is used for taking K training tuples with the minimum Euler distance, setting a label which appears most times in the K training tuples as a type of a test tuple, and dividing a training sample into at least one region according to the type of the test tuple;
the linear correlation program module is used for calculating the error rate and selecting a K value corresponding to the minimum error rate, wherein the K value corresponding to the minimum error rate indicates that the wind speeds of K wind generating sets in the wind power plant have linear correlation,
wherein the wind speed prediction program module comprises:
the model building program module is used for respectively carrying out linear regression analysis on the wind speed data in different areas and building a linear regression model according to the linear regression analysis result;
a wind speed prediction program module which constructs a loss function according to the linear regression model and the historical data, solves the corresponding parameter when the loss function is minimum, predicts the wind speed of a specific wind generating set in the wind power plant based on the solved parameter and the established linear regression model,
wherein the historical data comprises historical wind speed data and historical operating condition data.
7. The prediction system of claim 6, wherein the data handler module comprises:
the target function program module is used for carrying out centralized processing on the acquired historical data and solving a target function according to a centralized processing result;
the maximum characteristic value program module is used for forming a maximum solving problem by the target function and the constraint condition together, constructing a Lagrangian function and solving the maximum solving problem to obtain a maximum characteristic value;
and the dimension reduction processing program module is used for obtaining a dimension reduction processing result according to the maximum eigenvector corresponding to the maximum eigenvalue.
8. The prediction system of claim 6, wherein the loss function minimum is found by least squares and/or gradient descent.
9. The prediction system of claim 6, further comprising:
and the deviation testing program module is used for testing whether the predicted wind speed has deviation or not according to the wind power curve of the specific wind generating set.
10. The prediction system of claim 6, further comprising:
and the reinforcement learning program module is used for reinforcement learning of the established linear regression model.
11. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, causes the processor to carry out the prediction method according to any one of claims 1 to 5.
12. A computer device comprising a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, causes the processor to perform the prediction method of any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710474284.1A CN109102101B (en) | 2017-06-21 | 2017-06-21 | Wind speed prediction method and system for wind power plant |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710474284.1A CN109102101B (en) | 2017-06-21 | 2017-06-21 | Wind speed prediction method and system for wind power plant |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109102101A CN109102101A (en) | 2018-12-28 |
| CN109102101B true CN109102101B (en) | 2021-01-26 |
Family
ID=64796014
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710474284.1A Active CN109102101B (en) | 2017-06-21 | 2017-06-21 | Wind speed prediction method and system for wind power plant |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109102101B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109413149B (en) * | 2018-09-19 | 2021-05-18 | 上海哔哩哔哩科技有限公司 | Information distribution control method, system, server, and computer-readable storage medium |
| CN110188939B (en) * | 2019-05-24 | 2023-12-12 | 上海电气风电集团股份有限公司 | Wind power prediction method, system, equipment and storage medium of wind power plant |
| CN110222899B (en) * | 2019-06-12 | 2022-04-22 | 哈尔滨工业大学 | Cross-country skiing track wind speed field prediction method |
| CN110942182A (en) * | 2019-11-14 | 2020-03-31 | 国网福建省电力有限公司建设分公司 | Method for establishing typhoon prediction model based on support vector regression |
| CN112308322B (en) * | 2020-11-03 | 2023-11-24 | 华翔翔能科技股份有限公司 | Multi-wind-field space-time wind speed prediction method and device and electronic equipment |
| CN112502899B (en) * | 2020-11-30 | 2021-11-16 | 东方电气风电有限公司 | Consumption reduction method for wind generating set |
| CN113297728A (en) * | 2021-04-30 | 2021-08-24 | 东方电气风电有限公司 | Single-fan virtual wind speed calculation method and system based on wind field wind speed correlation |
| CN114048930B (en) * | 2022-01-13 | 2022-05-17 | 广东电网有限责任公司揭阳供电局 | Ultra-short-term wind power prediction method and device |
| CN115081677B (en) * | 2022-05-12 | 2024-02-02 | 暨南大学 | Wind speed prediction method and device, electronic equipment and readable storage medium |
| CN116070471B (en) * | 2023-04-06 | 2023-06-20 | 浙江远算科技有限公司 | Wind driven generator simulation acceleration method and system based on reduced order decomposition processing |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103324980A (en) * | 2013-04-25 | 2013-09-25 | 华北电力大学（保定） | Wind power station wind speed prediction method |
| CN104239964A (en) * | 2014-08-18 | 2014-12-24 | 华北电力大学 | Ultra-short-period wind speed prediction method based on spectral clustering type and genetic optimization extreme learning machine |
| CN105320809A (en) * | 2015-09-30 | 2016-02-10 | 西安电子科技大学 | Wind speed prediction method for wind farm spatial correlation |
| CN106067073A (en) * | 2016-05-30 | 2016-11-02 | 都城绿色能源有限公司 | A kind of wind power forecasting method based on wind-resources correlation analysis |
| CN106779201A (en) * | 2016-12-08 | 2017-05-31 | 贵州电网有限责任公司电力科学研究院 | A kind of wind speed modeling method of consideration plateau mountain area blower fan position |
| CN106875037A (en) * | 2016-12-30 | 2017-06-20 | 东软集团股份有限公司 | Wind-force Forecasting Methodology and device |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8930299B2 (en) * | 2010-12-15 | 2015-01-06 | Vaisala, Inc. | Systems and methods for wind forecasting and grid management |
-
2017
- 2017-06-21 CN CN201710474284.1A patent/CN109102101B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103324980A (en) * | 2013-04-25 | 2013-09-25 | 华北电力大学（保定） | Wind power station wind speed prediction method |
| CN104239964A (en) * | 2014-08-18 | 2014-12-24 | 华北电力大学 | Ultra-short-period wind speed prediction method based on spectral clustering type and genetic optimization extreme learning machine |
| CN105320809A (en) * | 2015-09-30 | 2016-02-10 | 西安电子科技大学 | Wind speed prediction method for wind farm spatial correlation |
| CN106067073A (en) * | 2016-05-30 | 2016-11-02 | 都城绿色能源有限公司 | A kind of wind power forecasting method based on wind-resources correlation analysis |
| CN106779201A (en) * | 2016-12-08 | 2017-05-31 | 贵州电网有限责任公司电力科学研究院 | A kind of wind speed modeling method of consideration plateau mountain area blower fan position |
| CN106875037A (en) * | 2016-12-30 | 2017-06-20 | 东软集团股份有限公司 | Wind-force Forecasting Methodology and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109102101A (en) | 2018-12-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109102101B (en) | Wind speed prediction method and system for wind power plant | |
| Wang et al. | Fault prognostics using dynamic wavelet neural networks | |
| US11840998B2 (en) | Hydraulic turbine cavitation acoustic signal identification method based on big data machine learning | |
| US20190227504A1 (en) | Computer System And Method For Monitoring Key Performance Indicators (KPIs) Online Using Time Series Pattern Model | |
| CN116380445B (en) | Equipment state diagnosis method and related device based on vibration waveform | |
| Costello et al. | Machine learning model for event-based prognostics in gas circulator condition monitoring | |
| CN110685868A (en) | Wind turbine generator fault detection method and device based on improved gradient elevator | |
| CN113049142A (en) | Temperature sensor alarm method, device, equipment and storage medium | |
| CN106650918B (en) | The method and apparatus for constructing system model | |
| US20200234165A1 (en) | Prediction method for aero-engine starting exhaust temperature | |
| RU2563161C2 (en) | Method and device of technical diagnostics of complex process equipment on basis of neuron net | |
| CN111898867B (en) | Airplane final assembly production line productivity prediction method based on deep neural network | |
| CN110210641A (en) | Wind direction prediction method and device for wind power plant | |
| CN116466672B (en) | Data center machine room parameter regulation and control method based on artificial intelligence and related device | |
| CN116086537A (en) | Equipment state monitoring method, device, equipment and storage medium | |
| Lutska et al. | Forecasting the efficiency of the control system of the technological object on the basis of neural networks | |
| CN114742363A (en) | Energy efficiency state evaluation method, system and medium for wind turbine generator | |
| CN113689072A (en) | Offshore wind turbine generator running state evaluation method based on Markov chain | |
| CN109683594A (en) | A kind of exceptional variable accurately identifies and localization method | |
| KR20200013146A (en) | Method for optimizing predictive algorithm based empirical model | |
| CN117077052A (en) | Dry-type transformer abnormality detection method based on working condition identification | |
| JP2005242803A (en) | Performance estimator, performance estimating method, and performance estimating program of machine | |
| US11339763B2 (en) | Method for windmill farm monitoring | |
| CN115795999A (en) | Performance abnormity early warning method for long-term service pumped storage unit | |
| CN114254834A (en) | Wind power plant wind speed prediction method based on machine learning |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |