CN110210641A - Wind direction prediction method and device for wind power plant - Google Patents
Wind direction prediction method and device for wind power plant Download PDFInfo
- Publication number
- CN110210641A CN110210641A CN201810167036.7A CN201810167036A CN110210641A CN 110210641 A CN110210641 A CN 110210641A CN 201810167036 A CN201810167036 A CN 201810167036A CN 110210641 A CN110210641 A CN 110210641A
- Authority
- CN
- China
- Prior art keywords
- wind
- wind direction
- nonlinear regression
- regression model
- nlrm
- 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 54
- 238000010219 correlation analysis Methods 0.000 claims abstract description 35
- 230000006870 function Effects 0.000 claims description 73
- 238000012360 testing method Methods 0.000 claims description 27
- 230000009977 dual effect Effects 0.000 claims description 19
- 238000000611 regression analysis Methods 0.000 claims description 13
- 238000004458 analytical method Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 10
- 238000007405 data analysis Methods 0.000 claims description 10
- 238000012549 training Methods 0.000 claims description 9
- 238000005457 optimization Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 230000005611 electricity Effects 0.000 claims description 7
- 238000012417 linear regression Methods 0.000 claims description 7
- 238000012544 monitoring process Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 4
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 238000004422 calculation algorithm Methods 0.000 description 16
- 230000008569 process Effects 0.000 description 10
- 230000008901 benefit Effects 0.000 description 5
- 238000010801 machine learning Methods 0.000 description 4
- 239000013598 vector Substances 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000013500 data storage Methods 0.000 description 2
- 238000000151 deposition Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002688 persistence Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- 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
-
- 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
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Data Mining & Analysis (AREA)
- Marketing (AREA)
- Mathematical Analysis (AREA)
- Operations Research (AREA)
- Health & Medical Sciences (AREA)
- Pure & Applied Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Mathematical Physics (AREA)
- Mathematical Optimization (AREA)
- Tourism & Hospitality (AREA)
- Computational Mathematics (AREA)
- Databases & Information Systems (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- General Engineering & Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Entrepreneurship & Innovation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Software Systems (AREA)
- Algebra (AREA)
- Probability & Statistics with Applications (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 present disclosure provides a wind direction prediction method and device for a wind farm, wherein the wind direction prediction method comprises the following steps: acquiring historical wind direction data of a wind power plant; analyzing the wind direction relativity of the acquired historical wind direction data; and determining a nonlinear regression model according to the wind direction correlation analysis result, and predicting the wind direction of the wind generation sets in the wind power plant based on the determined nonlinear regression model. The method can more accurately predict the wind directions of the wind generation sets aiming at different machine positions, and when a wind direction instrument of the wind generation set breaks down, wind direction compensation is carried out by using the predicted wind directions so that the wind generation set keeps higher wind efficiency.
Description
Technical field
The present invention relates to technical field of wind power, in particular, be related to a kind of wind direction prediction technique for wind power plant and
Wind direction prediction meanss.
Background technique
Due to Wind turbines high dry running outdoors, and the anemoscope of Wind turbines is placed in outside the hatch cover of Wind turbines,
The anemoscope running environment of Wind turbines is extremely complex: vibration, the various extreme conditions such as dust, be exposed to the sun, freeze, drenching with rain;Marine,
The installation unit of beach wind field also suffers salt air corrosion year in year out, therefore the probability that anemoscope breaks down is relatively
It is high.
Under normal conditions, the emergency control method taken when the prior art fails for the anemoscope of Wind turbines are as follows:
By Wind turbines all in wind field centered on respectively, three nearest Wind turbines of other geographical locations are passed through into SCADA system
System forms small network;When Wind turbines operate normally, Wind turbines are anti-by SCADA system by collected wind direction signals
It feeds central controller, the various data including wind direction of central controller real-time display Wind turbines;Work as Wind turbines
Anemoscope when breaking down, the anemoscope of failure carries out heterochromatic display in significant position by central controller, and to prison
Control personnel give a warning;Simultaneously, it is assumed that the wind direction of 3 adjacent Wind turbines is similar, by the anemoscope of similar Wind turbines
Data pass central controller back, after central controller confirms that the information of 1-2 platform Wind turbines is effective after screening, SCADA
System carries out selection according to priority and sends the Wind turbines of failure for the data of selection to make Wind turbines can
To obtain wind direction data continuous service after anemoscope failure.
However, since the prior art only assumes that the wind direction of adjacent Wind turbines is identical, do not consider certain distance it
Wind vector afterwards, it is very big for the prediction deviation of wind direction, it is difficult to meet the safe operation of Wind turbines, therefore, it is necessary to one kind
It more accurately predicts the wind direction of a certain Wind turbines and the Wind turbines that anemoscope breaks down is carried out with the side of wind direction compensation
Method and equipment.
Summary of the invention
To solve the above-mentioned problems and/or disadvantage, and at least advantages described below is provided, present disclose provides a kind of use
In the wind direction prediction technique and wind direction prediction meanss of wind power plant.
It is an aspect of the invention to provide a kind of wind direction prediction technique for wind power plant, the wind direction prediction technique packet
It includes following steps: obtaining the history wind direction data of wind power plant;Wind direction correlation analysis is carried out to the history wind direction data of acquisition;Root
Nonlinear regression model (NLRM) is determined according to wind direction correlation analysis result, and wind-powered electricity generation is predicted based on determining nonlinear regression model (NLRM)
The wind direction of Wind turbines in.
The step of carrying out wind direction correlation analysis to the history wind direction data of acquisition may include: to the history wind direction number
According to being pre-processed;Pretreated data are divided into trained tuple and test tuple, calculate separately test tuple and each instruction
Practice the distance between tuple;Choose it is described apart from the smallest K trained tuple, and by frequency of occurrence in described K trained tuple
Most labels is set as the classification of test tuple, and training tuple is divided at least one area according to the classification of test tuple
Domain.
The step of carrying out wind direction correlation analysis to the history wind direction data of acquisition can also include: calculating error rate, and
Select K value corresponding with minimal error rate, wherein K value corresponding with minimal error rate is expressed as K Wind turbines in wind power plant
Wind direction have non-linear dependencies.
It may include: respectively to not same district the step of determining nonlinear regression model (NLRM) according to wind direction correlation analysis result
History wind direction data in domain carries out nonlinear regression analysis, and is established according to the result of nonlinear regression analysis non-linear time
Return model;Cost function is constructed according to nonlinear regression model (NLRM) and history wind direction data;By drawing to the cost function of building
Enter slack variable to establish majorized function;Nonlinear regression model (NLRM) is determined by calculation optimization functional minimum value.
It may include: by foundation the step of determining nonlinear regression model (NLRM) by calculation optimization functional minimum value
Majorized function introduce Lagrangian and construct dual problem function;It is obtained by being calculated dual problem function
The equation of each variable in dual problem function;Acquisition is substituted into for the equation of each variable the majorized function established come
Obtain new majorized function;Nonlinear regression model (NLRM) is determined by calculating the minimum value of new majorized function.
The wind direction prediction technique can also include: to carry out intensified learning to determining nonlinear regression model (NLRM).
The wind direction prediction technique can also include: when the anemoscope of specific Wind turbines breaks down, based on determination
Nonlinear regression model (NLRM) predict the wind directions of the specific Wind turbines, and using the wind direction of prediction to the specific wind turbine
Group carries out wind direction compensation;The wind power curve of the specific Wind turbines is monitored to determine whether the specific Wind turbines continue
It remains operational.
The wind power curve of the specific Wind turbines is monitored to determine whether the specific Wind turbines continue to keep fortune
If capable step may include: to monitor that exception occurs in the wind power curve of the specific Wind turbines, to described specific
Wind turbines carry out shutdown processing;If monitoring that the wind power curve of the specific Wind turbines does not occur exception, use
The wind direction of prediction keeps the operations of the specific Wind turbines.
It is another aspect of the invention to provide a kind of wind direction prediction meanss for wind power plant, the wind direction prediction meanss
Include: data acquisition module, is configured as obtaining the history wind direction data of wind power plant;Data analysis module is configured as to obtaining
The history wind direction data taken carries out wind direction correlation analysis;Wind direction prediction module is configured as according to wind direction correlation analysis knot
Fruit determines nonlinear regression model (NLRM), and predicts the specific Wind turbines in wind power plant based on determining nonlinear regression model (NLRM)
Wind direction.
Data analysis module can be configured as: carry out in advance to the history wind direction data obtained by data acquisition module
Reason;Pretreated data are divided into trained tuple and test tuple, are calculated separately between test tuple and each trained tuple
Distance;Choose the label described apart from the smallest K trained tuple and frequency of occurrence in described K trained tuple is most
It is set as the classification of test tuple, training tuple is divided by least one region according to the classification of test tuple.
Data analysis module can be additionally configured to: error rate calculated, and selects K value corresponding with minimal error rate,
In, the wind direction that K value corresponding with minimal error rate is expressed as K Wind turbines in wind power plant has non-linear dependencies.
Wind direction prediction module can be configured as: carry out nonlinear regression to the history wind direction data in different zones respectively
Analysis, and nonlinear regression model (NLRM) is established according to the result of nonlinear regression analysis;According to nonlinear regression model (NLRM) and history
Wind direction data constructs cost function;Majorized function is established by introducing slack variable to the cost function of building;Pass through meter
The minimum value of majorized function is calculated to determine nonlinear regression model (NLRM).
Wind direction prediction module can be additionally configured to: be constructed by introducing Lagrangian to the majorized function of foundation
Dual problem function;The equation of each variable in dual problem function is obtained by being calculated dual problem function;
The equation for each variable of acquisition is substituted into the majorized function of foundation to obtain new majorized function;By calculating newly excellent
Change functional minimum value to determine nonlinear regression model (NLRM).
The wind direction prediction meanss can also include intensified learning module, be configured as to determining nonlinear regression model (NLRM)
Carry out intensified learning.
The wind direction prediction meanss can also include wind direction compensating module, be configured as: when the wind direction of specific Wind turbines
When instrument breaks down, using the wind direction for the specific Wind turbines predicted by determining nonlinear regression model (NLRM) to the specific wind
Motor group carries out wind direction compensation;The wind power curve of the specific Wind turbines is monitored whether to determine the specific Wind turbines
Continue to remain operational.
Wind direction compensating module can be additionally configured to: if monitoring that the wind power curve of the specific Wind turbines occurs
It is abnormal, then shutdown processing is carried out to the specific Wind turbines;If monitoring the wind power curve of the specific Wind turbines
Do not occur exception, then keeps the operation of the specific Wind turbines using the wind direction of prediction.
It is another aspect of the invention to provide a kind of computer readable storage mediums, are stored with computer program, special
Sign is, when the computer program is run by processor, processor executes above-described wind direction prediction technique.
It is another aspect of the invention to provide a kind of computer equipments, including processor and store depositing for computer program
Reservoir, which is characterized in that when the computer program is run by processor, processor executes above-described wind direction prediction side
Method.
The Wind turbines institute of different seat in the plane points can be directed to based on the wind direction prediction technique described above for wind power plant
Locate environment, predicts the wind direction of the Wind turbines of each seat in the plane point.Using correlation analysis and the recurrence based on machine learning divides
Analysis modeling, can predict the wind direction of some Wind turbines more accurately, it is made still may be used after anemoscope failure
To keep higher wind-powered electricity generation efficiency, it can be effectively reduced downtime as emergency system, improve the available of Wind turbines
Rate increases the unit output of Wind turbines, hence it is evident that increase economic efficiency.
Detailed description of the invention
Pass through the description carried out below in conjunction with attached drawing, above and other aspects, the feature of exemplary embodiment of the present invention
It will be more readily apparent from advantage, in the accompanying drawings:
Fig. 1 shows the flow chart of the prediction technique of the wind power plant wind direction of embodiment according to the present invention;
Fig. 2 shows use closest node (KNN) algorithms of embodiment according to the present invention to carry out showing for correlation analysis
Figure;
Fig. 3 shows the flow chart of the method for the determination nonlinear regression model (NLRM) of embodiment according to the present invention;
Fig. 4 shows the flow chart of the method for the compensation of wind power plant wind direction of embodiment according to the present invention;
Fig. 5 shows the block diagram of the wind direction prediction meanss for wind power plant of embodiment according to the present invention.
In the accompanying drawings, identical label will be understood to refer to identical element, feature and structure.
Specific embodiment
The present invention of the description to help comprehensive understanding to be defined by the claims and their equivalents referring to the drawings is provided
Exemplary embodiment.Description referring to the drawings includes various specific details to help to understand, but the specific detail
It only is seen as illustrative.Therefore, it will be appreciated by those of ordinary skill in the art that not departing from scope and spirit of the present invention
In the case where, the embodiments described herein can be made various changes and modifications.In addition, for clarity and briefly, public affairs can be omitted
Know the description of function and structure.
In the following, exemplary embodiment of the present invention is described in detail with reference to the attached drawings.It should be understood that exemplary reality according to the present invention
The wind direction prediction technique and wind direction prediction meanss for wind power plant for applying example can be applied to various Wind turbines, by being directed to
It the analysis of history wind direction data and is modeled using correlation analysis and the regression analysis based on machine learning, it can be more accurate
The wind direction of a certain Wind turbines is predicted on ground, it is made still can to keep higher wind-powered electricity generation efficiency after anemoscope failure,
Downtime can be effectively reduced as emergency system, the availability of wind-powered electricity generation blower is improved, increases the unit of Wind turbines
Output, so that economic benefit significantly improves.
Fig. 1 shows the flow chart of the prediction technique of the wind power plant wind direction of embodiment according to the present invention.
Referring to Fig.1, in step S101, the history wind direction data of wind power plant is obtained.According to an embodiment of the invention, obtain
History wind direction data may include that the history wind direction data of wind power plant is read from SCADA system.
In step S102, wind direction correlation analysis is carried out to the history wind direction data of acquisition.Firstly, will in step s101
The history wind direction data of acquisition is pre-processed, and then carries out wind direction correlation analysis to pretreated data.
In accordance with an embodiment of the present disclosure, closest node (KNN) algorithm can be used to carry out correlation analysis.It is general next
It says, it is not that whole Wind turbines in wind field all have correlation that the wind field area where Wind turbines is larger, may be due to
Landform reason makes a few Wind turbines have correlation on wind direction, in this way if calculating sky by establishing physical model
Air-flow mechanics difficulty is larger and different wind fields can only be calculated one by one.In the present embodiment, use is closest
Node (KNN) algorithm carries out correlation analysis, predicts being relatively close in wind field on wind direction by building mathematical model
Wind turbines.Wherein, KNN algorithm is mainly used in the identification to unknown things, that is, judges which classification is unknown things belong to;
KNN algorithm judges that thought is that the feature and which kind of other known things of unknown things are judged based on Euclidean axiom
Feature is closest.
After pre-processing to the history wind direction data of acquisition, pretreated data are divided into trained tuple and survey
Tuple is tried, test the distance between tuple and each trained tuple are calculated separately based on Euclidean axiom, wherein calculating two
When similarity between point, usually using Euclidean distance, equation can be indicated are as follows:
Wherein, xjIndicate test tuple, yjIndicate training tuple, deuc(x, y) is expressed as the distance of point-to-point transmission.
Then, choose it is calculated apart from the smallest K trained tuple according to equation (1), and by described K trained tuple
The most label of middle frequency of occurrence be set as test tuple classification, according to test tuple classification will training tuple be divided into
A few region.Error rate is finally calculated, and selects K value corresponding with minimal error rate, wherein is corresponding with minimal error rate
The wind direction that K value is expressed as K Wind turbines in wind power plant has non-linear dependencies.Root is described in detail below with reference to Fig. 2
The process of correlation analysis is carried out according to use closest node (KNN) algorithm of the embodiment of the present invention.
Fig. 2 shows use closest node (KNN) algorithms of embodiment according to the present invention to carry out showing for correlation analysis
Figure.
In Fig. 2,16 Wind turbines are chosen as training data, it are numbered respectively, input test data, count
Calculate Euler's distance between test data and each training data, the smallest K point of selected distance, class where K point before determining
Other frequency of occurrences, the highest classification of the frequency of occurrences is classified as the prediction of test data in K point before returning.It can be with from Fig. 2
Find out, the training data is divided into four regions after carrying out correlation analysis by KNN algorithm, wherein there is correlation
Number be 01,02,03,04,05,08 Wind turbines be classified as the same area, be 06 and 07 with the number of correlation
Wind turbines are classified as the same area, and the Wind turbines of the number 10,11,12,13 with correlation are classified as the same area, have
There are the Wind turbines of the number 09,14,15,16 of correlation to be classified as the same area.KNN algorithm is selected to carry out in the present embodiment
Correlation analysis, however, the disclosure is not limited to this algorithm.
Back to Fig. 1, after by step S102, wind field is had been divided as different regions.Each area divided
The wind vector of Wind turbines in domain all has correlation.In step S103, determined according to wind direction correlation analysis result
Nonlinear regression model (NLRM), and predict based on determining nonlinear regression model (NLRM) the wind direction of the specific Wind turbines in wind power plant.
The nonlinear regression model (NLRM) determined for predicting the wind direction of specific Wind turbines is described in detail below with reference to Fig. 3.
Fig. 3 shows the flow chart of the method for the determination nonlinear regression model (NLRM) of embodiment according to the present invention.According to this public affairs
Support vector regression (SVR) Lai Jianli nonlinear regression model (NLRM) can be used in the embodiment opened.
In step S301, nonlinear regression analysis is carried out to the history wind direction data in different zones respectively, and according to non-
The result of linear regression analysis establishes nonlinear regression model (NLRM).According to an embodiment of the invention, respectively to four divided in Fig. 2
Wind direction data in a different zones carries out nonlinear regression analysis, for example, to by numbering the wind for being 01,02,03,04,05,08
Wind direction data in the region of motor group composition carries out nonlinear regression analysis.According to Regression Analysis Result construct one it is non-thread
Property regression model function d (x).D (x) can be indicated in the form of d (x)=(ω x)+b, wherein x is expressed as history wind direction number
According to ω is expressed as hyperspace weight vectors, and b is expressed as intercept, and d (x) is expressed as the wind direction of prediction.In SVR algorithm, SVR
Purpose be that regression function determined by estimation ω and b.
In step S302, cost function is constructed according to the nonlinear regression model (NLRM) of foundation and history wind direction data.At this
In embodiment, according to SVR algorithm, cost function is expressed as one group of data to the distance of matched curve and is found in distance value
Maximum value.The cost function of SVR can indicate are as follows:
Cost (x)=max (0, | d (x)-g (x) |-ε) (2)
Wherein, | d (x)-g (x) | the wind direction data of a certain Wind turbines is indicated to the distance of plane of regression, and d (x) is expressed as
The wind direction of prediction, g (x) indicate practical wind direction value, and ε indicates tolerance value.Due to data it is not possible that all on plane of regression, and away from
From the sum of it is relatively large, therefore the distance of all data to plane of regression can give a tolerance value ε prevents over-fitting.The ginseng
Number is empirical parameter, is needed manually given.If the distance of data to plane of regression is less than ε, otherwise cost 0 is | d
(x)-g(x)|-ε。
In step S303, majorized function is established by introducing slack variable to the cost function of building.In the process,
Consider slack variable ξi、It is as follows to obtain constraint condition:
It can be seen that the above problem is converted into the problem of minimizing slack variable after introducing slack variable.This
In, it is assumed that the calculated result of ω meets normal distribution, according to Bayes's linear regression model (LRM), has L2 norm constraint to ω, thus
Establish majorized function:
Wherein, C indicates penalty factor, and C value is empirical parameter given by man.
Wherein it is possible to which peer-to-peer (4) is deformed, it is converted into equation (5):
In step S304, nonlinear regression model (NLRM) is determined by calculation optimization functional minimum value.It is built in step S303
After having found majorized function, dual problem function is constructed by introducing Lagrangian to the majorized function of foundation.Specifically,
By Lagrangian α, α*、β、β*It is introduced into majorized function, converts dual problem for optimization problem, wherein dual problem
Function can be indicated by equation (6):
By dual problem function respectively to variable ω, b, ξi、C seeks partial derivative, and then enabling the partial derivative sought is 0,
Obtain following equation (7) (8) (9) (10) (11):
Acquisition is directed to each variable ω, b, ξi、The equation of C, which substitutes into the majorized function (5) established, to be obtained newly
Majorized function, new majorized function can indicate with following equation (12):
It should be noted that during the equation for each variable obtained according to partial derivative is substituted into equation (5), variable
ξi、β、β*It has cancelled out each other in calculating process.
In addition, if then equation (12) can be deformed into equation (13) using kernel function:
After obtaining new majorized function, by the minimum for calculating new majorized function (i.e. equation (12) or equation (13))
Value determines nonlinear regression model (NLRM).Finally, the wind direction of wind power plant is predicted by the nonlinear regression model (NLRM) function after determination.
However, the disclosure is not limited to establish nonlinear regression model (NLRM) using SVR algorithm.
In addition, according to an embodiment of the invention, the invention also includes carry out extensive chemical to determining nonlinear regression model (NLRM)
It practises.Intensified learning is a kind of important machine learning method, and in the fields such as intelligent control and analysis prediction, there are many applications.Tool
Body, in the case where the anemoscope of Wind turbines operates normally, constantly collect, accumulation wind direction data, wind according to the present invention
Wind direction persistence forecasting and compare predicted value and actual value to prediction technique, with being continuously increased for data volume,
The loss function value of matched curve will be reduced constantly, make the wind direction prediction model in the process of running constantly enhancing algorithm study with
It is more nearly actual value, increases the accuracy of prediction.
Fig. 4 shows the flow chart of the method for the compensation of wind power plant wind direction of embodiment according to the present invention.
The history wind direction data of wind power plant is obtained in step S401 referring to Fig. 4.In step S402, to the history wind of acquisition
Wind direction correlation analysis is carried out to data.In the present embodiment, wind direction correlation analysis is carried out using KNN algorithm.It is calculated in KNN
In method, the neighbour's sample selected is all the object correctly classified, one on determining class decision only in accordance with arest neighbors or
The classification of several samples determines the generic of the sample that will be classified.For example, in Fig. 2, by the number with correlation
It is classified as the same area for 01,02,03,04,05,08 Wind turbines, i.e., these Wind turbines are determined as a classification.
In step S403, nonlinear regression model (NLRM) is determined according to wind direction correlation analysis result.Here, it is calculated according to SVR
Method determines nonlinear regression model (NLRM).The process of step S403 is identical as the process of step S103, in the present embodiment, not to this
It repeats again.
S405 is entered step when the anemoscope of a certain Wind turbines breaks down in step S404.Otherwise, the wind-powered electricity generation
Unit keeps current operating conditions to continue to run.
In step S405, the wind direction of the Wind turbines to break down is predicted based on determining nonlinear regression model (NLRM), and
Wind direction compensation is carried out to the Wind turbines using the wind direction of prediction, so that the Wind turbines remain operational.For example, if a certain wind
When the anemoscope of motor group breaks down, then the Wind turbines start according to the wind direction correlation analysis result in step S402
And the nonlinear regression model (NLRM) determined in step S403, the wind direction of the Wind turbines of anemoscope failure is predicted, it will be pre-
The wind direction of survey keeps the operation of the Wind turbines, i.e. progress wind direction compensation as the wind direction under current state.
Meanwhile the wind power curve of specific Wind turbines is monitored to determine whether specific Wind turbines continue to remain operational.
If exception occurs in the wind power curve monitored, shutdown processing is carried out to Wind turbines, to prevent safety accident occurs;If
The wind power curve monitored does not occur exception, then the operation of Wind turbines is kept using the wind direction of prediction, it can in wind
Wind turbines are made to continue to remain operational in the case where breaking down to instrument.This avoid anemoscope once damaging, by wind turbine
Group directly shut down handle and caused by generated energy and economic loss.
Fig. 5 shows the block diagram of the wind direction prediction meanss for wind power plant of embodiment according to the present invention.
Referring to Fig. 5, wind direction prediction meanss 500 include data acquisition module 501, are configured as obtaining the history wind of wind power plant
To data;Data analysis module 502 is configured as carrying out wind direction correlation analysis to the history wind direction data of acquisition;Wind direction is pre-
Module 503 is surveyed, is configured as determining nonlinear regression model (NLRM) according to wind direction correlation analysis result, and based on determining non-thread
Property regression model predicts the wind directions of the Wind turbines in wind power plant;Intensified learning module 504 is configured as to determining non-thread
Property regression model carry out intensified learning;And wind direction compensating module 505, it is configured as: when the anemoscope of a certain Wind turbines goes out
When existing failure, using the wind direction for the Wind turbines predicted by determining nonlinear regression model (NLRM) to the Wind turbines into
Row wind direction compensates so that the Wind turbines remain operational and monitor the wind power curve of the Wind turbines to determine the wind turbine
Whether group continues to remain operational.
During carrying out wind direction correlation analysis to the data of acquisition, data analysis module 502 is to by data acquisition
The history wind direction data that module 501 obtains is pre-processed;Pretreated data are divided into trained tuple and test tuple, point
The distance between tuple and each trained tuple Ji Suan not tested;Choose it is described apart from the smallest K trained tuple, and will be described
The most label of frequency of occurrence is set as the classification of test tuple in K trained tuple, will be trained according to the classification of test tuple
Tuple is divided at least one region.
After test tuple set is completed, data analysis module 502 calculates error rate, and selects and minimal error rate
Corresponding K value, wherein the wind direction that K value corresponding with minimal error rate is expressed as K Wind turbines in wind power plant has non-linear
Correlation.It should be noted that this process can carry out the correlation analysis of wind direction by the way of step S102, it is no longer superfluous here
It states, but the disclosure and the unlimited method using step S102.
After to wind direction correlation analysis, wind direction prediction module 503 is respectively to being divided by data analysis module 502
History wind direction data in different zones carries out nonlinear regression analysis, and is determined according to the result of nonlinear regression analysis non-
Linear regression model (LRM) constructs cost function according to nonlinear regression model (NLRM) and history wind direction data, passes through the cost to building
Function introduces slack variable to establish majorized function, determines nonlinear regression model (NLRM) by calculation optimization functional minimum value.
In determining nonlinear regression model (NLRM), wind direction prediction module 503 by the majorized function of foundation introduce Lagrangian come
Construct dual problem function, obtained by being calculated dual problem function each variable in dual problem function etc.
The equation for each variable of acquisition is substituted into the majorized function of foundation to obtain new majorized function, by calculating newly by formula
The minimum value of majorized function determine nonlinear regression model (NLRM).This process is identical as the process in step S103, here no longer
It is described in detail.
When the anemoscope of a certain Wind turbines breaks down, 505 use of wind direction compensating module is by wind direction prediction module institute
Current wind direction of the wind direction for the Wind turbines of prediction as the Wind turbines carries out wind direction compensation to the Wind turbines,
Whether the wind power curve for detecting the Wind turbines simultaneously is normal, if it is different to monitor that the wind power curve of the Wind turbines occurs
Often, then shutdown processing is carried out to the Wind turbines;If monitoring that the wind power curve of the Wind turbines does not occur exception, make
The operation of the Wind turbines is kept with the wind direction of prediction.
In addition, according to an embodiment of the invention, intensified learning module can also be passed through when Wind turbines operate normally
504 pairs of determining nonlinear regression model (NLRM)s constantly carry out intensified learning, are used in the pre- nonlinear regression model (NLRM) to determine the wind direction and exist
Constantly enhance algorithm in operational process to learn so that the wind direction of prediction is more nearly actual value.
A kind of wind direction prediction technique and device for wind power plant of embodiment according to the present invention, the prediction technique pass through
Analysis and the modeling of wind power plant wind direction to history wind direction data, are built using correlation analysis and the regression analysis based on machine learning
Mould more can accurately predict the wind direction of some specific wind power generating set, it is made still may be used after anemoscope failure
To keep higher wind-powered electricity generation efficiency, downtime can be effectively reduced as emergency system, improving wind power generating set can benefit
With rate, increase wind power generating set unit output, economic benefit is obvious.
A kind of wind direction prediction technique for wind power plant of embodiment according to the present invention can be realized as computer-readable note
Computer-readable code on recording medium, or can be sent by transmission medium.Computer readable recording medium is can to store
Hereafter the arbitrary data storage device for the data that can be read by computer system.Computer-readable recording medium storage has computer
Program, when which is run by processor, processor executes the prediction technique of the wind power plant wind direction in the disclosure.It calculates
The example of machine readable medium recording program performing includes read-only memory (ROM), random access memory (RAM), CD (CD)-ROM, number
Versatile disc (DVD), tape, floppy disk, optical data storage device, but not limited to this.Transmission medium may include by network or
The carrier wave that various types of communication channels are sent.Computer readable recording medium also can be distributed in the department of computer science of connection network
System, so that computer-readable code is stored and executed in a distributed fashion.
Another embodiment of the present invention provides a kind of computer equipment, including processor and stores depositing for computer program
Reservoir, when the computer program is run by processor, processor executes the prediction technique of the wind power plant wind direction in the disclosure.
Although the present invention, art technology has been shown and described referring to certain exemplary embodiments of the invention
Personnel will be understood that, can be into the case where not departing from the spirit and scope of the present invention being defined by the claims and their equivalents
Various changes on row various forms and details.
Claims (18)
1. a kind of wind direction prediction technique for wind power plant, which is characterized in that the wind direction prediction technique the following steps are included:
Obtain the history wind direction data of wind power plant;
Wind direction correlation analysis is carried out to the history wind direction data of acquisition;
Nonlinear regression model (NLRM) is determined according to wind direction correlation analysis result, and based on determining nonlinear regression model (NLRM) come pre-
Survey the wind direction of the Wind turbines in wind power plant.
2. wind direction prediction technique as described in claim 1, which is characterized in that carry out wind direction phase to the history wind direction data of acquisition
Close property analysis the step of include:
The history wind direction data is pre-processed;
Pretreated data are divided into trained tuple and test tuple, are calculated separately between test tuple and each trained tuple
Distance;
Choose the label setting described apart from the smallest K trained tuple and frequency of occurrence in described K trained tuple is most
For the classification for testing tuple, training tuple is divided by least one region according to the classification of test tuple.
3. wind direction prediction technique as claimed in claim 2, which is characterized in that carry out wind direction phase to the history wind direction data of acquisition
The step of closing property analysis further include:
Error rate is calculated, and selects K value corresponding with minimal error rate,
Wherein, the wind direction that K value corresponding with minimal error rate is expressed as K Wind turbines in wind power plant has nonlinear correlation
Property.
4. wind direction prediction technique as described in claim 1, which is characterized in that determined according to wind direction correlation analysis result non-
The step of linear regression model (LRM) includes:
Nonlinear regression analysis is carried out to the history wind direction data at least one described region respectively, and according to nonlinear regression
The result of analysis establishes nonlinear regression model (NLRM);
Cost function is constructed according to nonlinear regression model (NLRM) and the history wind direction data;
Majorized function is established by introducing slack variable to the cost function of building;
Nonlinear regression model (NLRM) is determined by calculation optimization functional minimum value.
5. wind direction prediction technique as claimed in claim 4, which is characterized in that determined by calculation optimization functional minimum value
The step of nonlinear regression model (NLRM) includes:
Dual problem function is constructed by introducing Lagrangian to the majorized function of foundation;
The equation of each variable in dual problem function is obtained by being calculated dual problem function;
The equation of each variable of acquisition is substituted into the majorized function of foundation to obtain new majorized function;
Nonlinear regression model (NLRM) is determined by calculating the minimum value of new majorized function.
6. wind direction prediction technique as described in claim 1, which is characterized in that further include: to determining nonlinear regression model (NLRM)
Carry out intensified learning.
7. wind direction prediction technique as described in claim 1, which is characterized in that further include:
When the anemoscope of specific Wind turbines breaks down, the specific wind is predicted based on determining nonlinear regression model (NLRM)
The wind direction of motor group, and wind direction compensation is carried out to the specific Wind turbines using the wind direction of prediction;And
The wind power curve of the specific Wind turbines is monitored to determine whether the specific Wind turbines continue to remain operational.
8. wind direction prediction technique as claimed in claim 7, which is characterized in that the wind power curve for monitoring specific Wind turbines is come
Determine whether the specific Wind turbines continue the step of remaining operational and include:
If it is abnormal to monitor that the wind power curve of the specific Wind turbines occurs, the specific Wind turbines are stopped
Machine processing;
If monitoring that the wind power curve of the specific Wind turbines does not occur exception, keep special using the wind direction of prediction
It is described to determine running of wind generating set.
9. a kind of wind direction prediction meanss for wind power plant, which is characterized in that the wind direction prediction meanss include:
Data acquisition module is configured as obtaining the history wind direction data of wind power plant;
Data analysis module is configured as carrying out wind direction correlation analysis to the history wind direction data of acquisition;
Wind direction prediction module is configured as determining nonlinear regression model (NLRM) according to wind direction correlation analysis result, and based on true
Fixed nonlinear regression model (NLRM) predicts the wind directions of the Wind turbines in wind power plant.
10. wind direction prediction meanss as claimed in claim 9, which is characterized in that data analysis module is configured as:
The history wind direction data obtained by data acquisition module is pre-processed;
Pretreated data are divided into trained tuple and test tuple, are calculated separately between test tuple and each trained tuple
Distance;
Choose the label setting described apart from the smallest K trained tuple and frequency of occurrence in described K trained tuple is most
For the classification for testing tuple, training tuple is divided by least one region according to the classification of test tuple.
11. wind direction prediction meanss as claimed in claim 10, which is characterized in that data analysis module is also configured to
Error rate is calculated, and selects K value corresponding with minimal error rate,
Wherein, the wind direction that K value corresponding with minimal error rate is expressed as K Wind turbines in wind power plant has nonlinear correlation
Property.
12. wind direction prediction meanss as claimed in claim 9, which is characterized in that wind direction prediction module is configured as:
Nonlinear regression analysis is carried out to the history wind direction data at least one described region respectively, and according to nonlinear regression
The result of analysis establishes nonlinear regression model (NLRM);
Cost function is constructed according to nonlinear regression model (NLRM) and history wind direction data;
Majorized function is established by introducing slack variable to the cost function of building;
Nonlinear regression model (NLRM) is determined by calculation optimization functional minimum value.
13. wind direction prediction meanss as claimed in claim 12, which is characterized in that wind direction prediction module is also configured to
Dual problem function is constructed by introducing Lagrangian to the majorized function of foundation;
The equation of each variable in dual problem function is obtained by being calculated dual problem function;
The equation of each variable of acquisition is substituted into the majorized function of foundation to obtain new majorized function;
Nonlinear regression model (NLRM) is determined by calculating the minimum value of new majorized function.
14. wind direction prediction meanss as claimed in claim 9, which is characterized in that further include intensified learning module, be configured as pair
Determining nonlinear regression model (NLRM) carries out intensified learning.
15. wind direction prediction meanss as claimed in claim 9, which is characterized in that further include wind direction compensating module, be configured as:
When the anemoscope of specific Wind turbines breaks down, the specific wind-powered electricity generation by determining nonlinear regression model (NLRM) prediction is used
The wind direction of unit carries out wind direction compensation to the specific Wind turbines;
The wind power curve of the specific Wind turbines is monitored to determine whether the specific Wind turbines continue to remain operational.
16. wind direction prediction meanss as claimed in claim 15, which is characterized in that wind direction compensating module is also configured to
If it is abnormal to monitor that the wind power curve of the specific Wind turbines occurs, the specific Wind turbines are stopped
Machine processing;
If monitoring that the wind power curve of the specific Wind turbines does not occur exception, using the wind direction of prediction to keep
State specific running of wind generating set.
17. a kind of computer readable storage medium, is stored with computer program, which is characterized in that the computer program is located
When managing device operation, processor perform claim requires wind direction prediction technique described in any one of 1-8.
18. a kind of computer equipment, the memory including processor and storage computer program, which is characterized in that the calculating
When machine program is run by processor, processor executes the wind direction prediction technique as described in any one of claim 1-8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810167036.7A CN110210641A (en) | 2018-02-28 | 2018-02-28 | Wind direction prediction method and device for wind power plant |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810167036.7A CN110210641A (en) | 2018-02-28 | 2018-02-28 | Wind direction prediction method and device for wind power plant |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110210641A true CN110210641A (en) | 2019-09-06 |
Family
ID=67778892
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810167036.7A Pending CN110210641A (en) | 2018-02-28 | 2018-02-28 | Wind direction prediction method and device for wind power plant |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110210641A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
CN111396248A (en) * | 2020-03-16 | 2020-07-10 | 明阳智慧能源集团股份公司 | Wind turbine generator set intelligent yaw control method based on short-term wind direction prediction |
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 |
CN116954169A (en) * | 2023-07-19 | 2023-10-27 | 天津市易控科技发展有限公司 | Information security control method, device, equipment and medium based on DCS control |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20130089738A (en) * | 2011-12-30 | 2013-08-13 | 주식회사 효성 | Predicting system for energy management system |
CN103401236A (en) * | 2013-07-30 | 2013-11-20 | 华北电力大学 | Wind power farm generator unit grouping method based on flow correlation of wind power farm |
CN105065201A (en) * | 2015-07-22 | 2015-11-18 | 天津瑞源电气有限公司 | Control method for ensuring normal running after damage of wind indicator of wind turbine unit |
CN105654207A (en) * | 2016-01-07 | 2016-06-08 | 国网辽宁省电力有限公司锦州供电公司 | Wind power prediction method based on wind speed information and wind direction information |
-
2018
- 2018-02-28 CN CN201810167036.7A patent/CN110210641A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20130089738A (en) * | 2011-12-30 | 2013-08-13 | 주식회사 효성 | Predicting system for energy management system |
CN103401236A (en) * | 2013-07-30 | 2013-11-20 | 华北电力大学 | Wind power farm generator unit grouping method based on flow correlation of wind power farm |
CN105065201A (en) * | 2015-07-22 | 2015-11-18 | 天津瑞源电气有限公司 | Control method for ensuring normal running after damage of wind indicator of wind turbine unit |
CN105654207A (en) * | 2016-01-07 | 2016-06-08 | 国网辽宁省电力有限公司锦州供电公司 | Wind power prediction method based on wind speed information and wind direction information |
Non-Patent Citations (3)
Title |
---|
王印松等: "一种基于相邻风机测量数据相关性分析的风速预测方法", 《华北电力大学学报(自然科学版)》 * |
王慧勤: "基于支持向量机的短期风速预测方法研究", 《中国优秀硕士学位论文全文数据库 基础科学辑(月刊)》 * |
许炳坤: "基于聚类算法和风向预测的风机偏航控制优化策略", 《基于聚类算法和风向预测的风机偏航控制优化策略》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
CN111396248A (en) * | 2020-03-16 | 2020-07-10 | 明阳智慧能源集团股份公司 | Wind turbine generator set intelligent yaw control method based on short-term wind direction prediction |
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 |
CN116954169A (en) * | 2023-07-19 | 2023-10-27 | 天津市易控科技发展有限公司 | Information security control method, device, equipment and medium based on DCS control |
CN116954169B (en) * | 2023-07-19 | 2024-06-07 | 天津市易控科技发展有限公司 | Information security control method, device, equipment and medium based on DCS control |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109102101B (en) | Wind speed prediction method and system for wind power plant | |
Smarsly et al. | Machine learning techniques for structural health monitoring | |
US8606418B1 (en) | Wind prediction for wind farms through the use of weather radar | |
CN110210641A (en) | Wind direction prediction method and device for wind power plant | |
CN115034248A (en) | Automatic diagnostic method, system and storage medium for equipment | |
CN109034191A (en) | One-dimensional telemetry exception interpretation method based on ELM | |
CN112508442A (en) | Transient stability evaluation method and system based on automation and interpretable machine learning | |
Jiang et al. | Short-term distribution system state forecast based on optimal synchrophasor sensor placement and extreme learning machine | |
CN110709789A (en) | Method and apparatus for monitoring the condition of a subsystem within a renewable power generation device or microgrid | |
Ferreira et al. | Probabilistic ramp detection and forecasting for wind power prediction | |
Khan et al. | AI based real-time signal reconstruction for wind farm with SCADA sensor failure | |
JP6373682B2 (en) | Wind power generation equipment stress estimation device, wind power generation equipment stress estimation method, wind power generation system | |
Lutska et al. | Forecasting the efficiency of the control system of the technological object on the basis of neural networks | |
Jiang et al. | Remaining useful life prediction of rolling bearings based on Bayesian neural network and uncertainty quantification | |
CN112580254B (en) | Multi-dimensional analysis and early warning method for converter transformer based on multi-level decision fusion | |
Zhang et al. | Condition based maintenance and operation of wind turbines | |
Fahim et al. | Data-driven techniques for optimizing the renewable energy systems operations | |
ȚIBOACĂ et al. | Design of short-term wind production forecasting model using machine learning algorithms | |
CN113627621B (en) | Active learning method for optical network transmission quality regression estimation | |
Sun et al. | A data-driven framework for tunnel infrastructure maintenance | |
KR102570248B1 (en) | System and Method for Photovoltaics Power Prediction | |
Tidriri et al. | Data-driven decision-making methodology for prognostic and health management of wind turbines | |
Tan et al. | Real-time life prediction of equipment based on optimized ARMA model | |
Bouazzi et al. | Deep recurrent neural networks based Bayesian optimization for fault diagnosis of uncertain GCPV systems depending on outdoor condition variation | |
Alamaniotis et al. | Optimal assembly of support vector regressors with application to system monitoring |
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 |