CN111476402A - Wind power generation capacity prediction method coupling meteorological information and EMD technology - Google Patents
Wind power generation capacity prediction method coupling meteorological information and EMD technology Download PDFInfo
- Publication number
- CN111476402A CN111476402A CN202010182883.8A CN202010182883A CN111476402A CN 111476402 A CN111476402 A CN 111476402A CN 202010182883 A CN202010182883 A CN 202010182883A CN 111476402 A CN111476402 A CN 111476402A
- Authority
- CN
- China
- Prior art keywords
- wind power
- data
- wind speed
- prediction
- wind
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 123
- 238000010248 power generation Methods 0.000 title claims abstract description 52
- 230000008878 coupling Effects 0.000 title claims abstract description 21
- 238000010168 coupling process Methods 0.000 title claims abstract description 21
- 238000005859 coupling reaction Methods 0.000 title claims abstract description 21
- 238000005516 engineering process Methods 0.000 title claims abstract description 17
- 230000006870 function Effects 0.000 claims abstract description 61
- 238000012549 training Methods 0.000 claims abstract description 58
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 33
- 238000012545 processing Methods 0.000 claims abstract description 18
- 230000002159 abnormal effect Effects 0.000 claims description 31
- 238000013528 artificial neural network Methods 0.000 claims description 26
- 230000008859 change Effects 0.000 claims description 25
- 238000004140 cleaning Methods 0.000 claims description 19
- 238000004364 calculation method Methods 0.000 claims description 18
- 238000003062 neural network model Methods 0.000 claims description 15
- 210000002569 neuron Anatomy 0.000 claims description 11
- 238000004891 communication Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 8
- 238000007620 mathematical function Methods 0.000 claims description 3
- 238000004422 calculation algorithm Methods 0.000 abstract description 12
- 238000011160 research Methods 0.000 abstract description 5
- 238000005457 optimization Methods 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 20
- 238000010586 diagram Methods 0.000 description 5
- 238000007781 pre-processing Methods 0.000 description 5
- 235000010627 Phaseolus vulgaris Nutrition 0.000 description 3
- 244000046052 Phaseolus vulgaris Species 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 238000009825 accumulation Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000035772 mutation Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000013179 statistical model Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000008570 general process Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 210000004205 output neuron Anatomy 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000002407 reforming Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000007704 transition Effects 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
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- 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
- 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)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Mathematical Analysis (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Pure & Applied Mathematics (AREA)
- Biophysics (AREA)
- Mathematical Optimization (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Computational Mathematics (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Evolutionary Biology (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Quality & Reliability (AREA)
- Water Supply & Treatment (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Probability & Statistics with Applications (AREA)
Abstract
The invention relates to a wind power generation capacity prediction method for coupling meteorological information and an EMD data processing technology, and belongs to the technical field of wind power generation capacity prediction. According to the method, firstly, the optimization processing of the historical wind power data distortion data is realized through a variable point method-quartile method, and the data quality of the original data series of the generated training set is improved. Secondly, the processed wind power series are decomposed through an empirical mode decomposition algorithm, and BP modeling training is carried out on the eigenmode functions with stronger regularity and stability after decomposition and meteorological information of corresponding time sequences so as to improve the final wind power prediction precision. The method can be used for predicting the generating capacity of a single wind power station at the middle-term time scale, the accuracy of the prediction result is better than that of a general BP modeling prediction model in the current research field, and the method is feasible and usable.
Description
Technical Field
The invention belongs to the technical field of wind power generation capacity prediction, and particularly relates to a wind power generation capacity prediction method based on a coupling meteorological information and EMD data processing technology.
Background
At present, the demand of our country for the clean energy consumption capacity is increasing day by day, and a series of problems are also provided for how to further bring a larger capacity of clean energy power into a power grid and effectively cooperate with the traditional power supply for operation and scheduling. Wind power, photovoltaic and the like belong to common clean energy, the output process of the wind power, photovoltaic and the like is influenced by meteorological conditions of the place where the power plant is located, the wind power, photovoltaic and the like often have the characteristic of intermittence, and certain obstacles are caused when the wind power, photovoltaic and the like are brought into a local power grid and a power generation plan is arranged in a unified mode. Based on the actual situation, the method has important significance for the accurate evaluation of the long, medium and short-term power generation capacity of the wind power station under each time scale and the safe and stable operation of a clean energy power grid. At present, the prediction of the power generation capacity of the wind power station at home and abroad has the following solving directions: (1) based on the traditional statistical model, starting from historical power generation data of the wind power station, the similar power generation data of particles in a certain time is subjected to extension in the meaning of the mathematical statistical model, so that the purpose of prediction is achieved. The basis of the prediction mode is that the wind condition of the wind power plant has certain inertia in time, and the inertia is stable for a period of time when the prediction scale is short. Therefore, based on the thought, point prediction or probability confidence interval prediction of the wind power plant can be realized by applying a Markov chain, an autoregressive moving average model and the like. However, for the prediction of the power generation capacity of the wind power station on the medium-long scale, the method based on state transition and linear regression may not be suitable any more at this time because the stationarity of the wind power change process is weakened. (2) And predicting the generating capacity of the wind power station based on a soft computing method. The method comprises the steps of deeply searching high-dimensional characteristics implicit in data, using a calculation method comprising a neural network, support vector regression, regression trees and the like, preprocessing historical data of the wind power station needing to predict the power generation capacity according to a required time step and using the preprocessed historical data as input conditions (such as wind speed, wind power and the like of the wind power station), using predicted single-step or multi-step wind speed or wind power as output of a model, simulating a dynamic change process of physical quantity through a black box model established in a soft calculation mode, and finally obtaining a prediction result. However, the mapping relation between the input condition and the output wind power of the method is a black box model, and the physical significance of the method is unclear, so that the method is not accurate enough in modeling expression of the corresponding relation between data under the condition that certain high-dimensional features are discontinuous (for example, some natural conditions affecting the wind power are changed and cannot be quantized), and further the prediction precision of the method is low under the condition. Therefore, how to overcome the defects of the prior art is a problem which needs to be solved urgently in the technical field of wind power generation capacity prediction at present.
Disclosure of Invention
The invention aims to solve the defects of the prior art and provides a wind power generation capacity prediction method for coupling meteorological information and an EMD data processing technology.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
the wind power generation capacity prediction method coupling meteorological information and an EMD technology comprises the following steps:
step (1), basic data required by a model is established, wherein the basic data comprises daily average output of a wind power station for more than 2 years, installed capacity of the wind power station and daily wind speed time sequence data, and the length of the data sequence is set to be N days;
step (2), carrying out data cleaning treatment on the historical daily average output of the wind power station in combination with historical wind speed data by adopting a variable point method-quartile method, removing invalid abnormal information caused by communication faults, artificial wind abandon and data recording errors, and obtaining a wind speed-daily average output curve of the wind power station; for the day with abnormal wind power, linear interpolation is carried out by adopting a linear interpolation method to obtain daily average output, and finally, a wind speed-wind power time sequence data set W { (v) which can be used for training is obtained1,p1),(v2,p2),…,(vi,pi),…,(vN,pN) Where v is 1,2, …, NiWind speed at day i, piIs the wind power on day i;
step (3), p in the sequence W1,p2,…,pnPerforming sequence decomposition on part of the I eigen-mode functions by adopting an empirical mode decomposition method to obtain corresponding I eigen-mode functions and a remainder;
step (4), for the 2 nd eigenmode function to the I th eigenmode function and the residual expression, i.e. IMF2、……、IMFI、R0The corresponding mathematical function of which is expressed as f1(vi),f2(vi),…,fI(vi) It is combined with the sequence WAre respectively formed into T1、……、TII new input training set Tm={vi,fm(vi) 1,2, …, I; determining a forecast period T for model predictionsforeSetting the number of neurons in the input layer of the training model corresponding to each eigenmode function as TforeThe number of neurons in the output layer is 1, and T is set for each training setiRespectively carrying out BP neural network training; finally, obtaining a wind speed-wind power eigenmode function BP neural network model suitable for power station wind power prediction;
step (5), predicting the day-ahead T of the starting point according to the input condition required in the model training prediction in the step (4)foreInputting the historical wind speed sequence of the day into the trained model (4), and predicting backward T from the prediction starting point by the model day by dayforeAnd (4) average output of the wind power station in each day, namely the predicted wind power generation capacity in the period to be predicted of the wind power station.
Further, it is preferable that the specific method of the step (2) is:
step (2.1), grouping the historical daily average output data of the wind power station and the corresponding wind speed data from small to large according to the wind speed interval of every 1m/S, and dividing the data into S1、S2、…、Sk、…SLL groups are provided, and the data length in each group is Dk;
Step (2.2), for each group Sk(k-1, 2, …, L) according to the formulaCalculate the variance q of each data pointtWherein p istThe average output force on the day of the t day,the mean of the average forces over all days in the group was followed by calculation of the rate of change of variance Δ for each data pointt=|qt-qt-1|;
Step (2.3), the variance change rate delta is processed by adopting a least square methodtIdentifying the change point, setting deltatObeying a two-segment linear model(ii) a By DeltatThe sum of the squares of the differences between the observed value and the theoretical value of (A) is used as an objective function, and the time or position of the point where the sum of the squares of the differences reaches the minimum value is used as an estimate of the position of the change point of the data series, to determine the S for each groupkThe data sequence of (1) is changed into a point sequence number, and partial abnormal values after the sequence number are discarded;
step (2.3), cleaning the numerical value points corresponding to the average output of the abnormal days in the wind speed interval after the abnormal interval is abandoned by adopting a quartile method; the daily average output interval [ P ] in the wind speed interval processed by the variable point methodmin,Pmax]Dividing the average into four parts, and recording the values at the positions of three division points as Q1、Q2、Q3(ii) a By a quarter of a bit distance IQR=Q3-Q1Determining an inner limit [ F ] of the daily average output value1,Fu]=[Q1-1.5IQR,Q3+1.5IQR](ii) a And determining the wind speed-daily average output data points outside the inner limit as daily average output abnormal data points for cleaning.
Further, it is preferable that in the step (4), the training step of the BP neural network training is performed by taking T ahead from the prediction start point day at ①foreInputting the daily wind speed data into the input layer of each eigenmode function BP neural network, ② obtaining the predicted value p of the 1 st daily average output of the predicted section from the output layerf1Corresponding eigenmode function values; adding the eigenmode function values to obtain pf1Obtaining a corresponding wind speed value v from a wind speed-daily average output curve f1③ moving the prediction starting point by 1 day, predicting the 2 nd, … … th and Tth according to the steps of ① and ②foreAnd (4) predicting the value.
The invention also provides a wind power generation capacity prediction system coupling meteorological information and an EMD technology, which comprises:
the data acquisition module is used for acquiring daily average output of the wind power station, installed capacity of the wind power station and daily wind speed time sequence data for more than 2 years;
the first processing module is used for cleaning the acquired data by adopting a variable point method-quartile method, and performing linear interpolation on the wind power abnormal day and the wind power abnormal day by adopting a linear interpolation method to obtain daily average output so as to finally obtain a wind speed-wind power time sequence data set for training;
the second processing module is used for performing sequence decomposition on the wind power part in the wind speed-wind power time sequence data set by adopting an empirical mode decomposition method, and performing BP neural network training on an input training set formed by the 2 nd eigenmode function, the I th eigenmode function, the remainder expression and the wind speed in the wind speed-wind power time sequence data set to obtain a BP neural network model;
and the wind power generation capacity prediction module is used for predicting the wind power generation capacity in the period to be predicted by adopting a BP neural network model.
The invention also provides electronic equipment which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, and is characterized in that the processor executes the program to realize the steps of the wind power generation capacity prediction method for coupling the meteorological information and the EMD technology.
The present invention additionally provides a non-transitory computer readable storage medium having stored thereon a computer program, characterized in that the computer program, when executed by a processor, implements the steps of the wind power generation capability prediction method as described above for coupling meteorological information with EMD techniques.
Aiming at the problems that objective distortion exists in historical data in wind power prediction, randomness and volatility of an output process are high, and high-dimensional characteristics are difficult to accurately find even though modeling is performed through a soft computing method, so that accurate prediction is realized, the wind power prediction is performed through a mode of two-step data preprocessing and neural network for modeling training of a processed data set, so that the problems are solved. Firstly, the optimization processing of the historical wind power data distortion data is realized through a variable point method-quartile method, and the data quality of the original data series of the generated training set is improved. Secondly, the processed wind power series are decomposed through an empirical mode decomposition algorithm, and BP modeling training is carried out on the eigenmode functions with stronger regularity and stability after decomposition and meteorological information of corresponding time sequences so as to improve the final wind power prediction precision. The method can be used for predicting the generating capacity of a single wind power station at the medium-term time scale (7-10d), the accuracy of the prediction result is superior to that of a general BP modeling prediction model in the current research field, and the method is feasible and usable.
Compared with the prior art, the invention has the beneficial effects that:
the invention relates to a method for forecasting and modeling the medium-term power generation capacity of a wind power plant by combining meteorological information and various data preprocessing methods. By preprocessing the data in two stages, high-dimensional features between corresponding data sets are easier to capture by a neural network algorithm, the speed of generating the model is higher, and the calculation accuracy is better.
According to the method, reasonable meteorological data are integrated, the historical wind power data of the wind power station are preprocessed by adopting a proper algorithm, the meteorological data of the wind power station and the historical wind power data of the station are integrated, and then a soft computing method is adopted to explore high-dimensional characteristics of the data to perform prediction computation, so that the wind power prediction problem is effectively solved; the method has the advantages that the physical influence factors of wind power generation are considered while the statistical analysis and soft computing method modeling are carried out on the historical time sequence data, the prediction precision is greatly improved compared with that of a method which uses one of the historical time sequence data and the soft computing method singly, the prediction requirement on the wind power generation capacity in the process of making a power grid scheduling plan is better met, and the method has important use and popularization values.
Drawings
FIG. 1 is a general process block diagram of the process of the present invention;
FIG. 2 is a single wind farm wind power series EMD decomposition diagram;
FIG. 3 is a comparison of the predicted results of the bean grinding mountain wind farm with the actual wind power process;
FIG. 4 is a schematic structural diagram of a wind power generation capability prediction system of the present invention coupling meteorological information and EMD technology;
fig. 5 is a schematic structural diagram of an electronic device according to the present invention.
Detailed Description
The present invention will be described in further detail with reference to examples.
It will be appreciated by those skilled in the art that the following examples are illustrative of the invention only and should not be taken as limiting the scope of the invention. The examples do not specify particular techniques or conditions, and are performed according to the techniques or conditions described in the literature in the art or according to the product specifications. The materials or equipment used are not indicated by manufacturers, and all are conventional products available by purchase.
The wind power generation capacity prediction method coupling meteorological information and an EMD technology comprises the following steps:
step (1), basic data required by a model is established, wherein the basic data comprises daily average output of a wind power station for more than 2 years, installed capacity of the wind power station and daily wind speed time sequence data, and the length of the data sequence is set to be N days;
step (2), carrying out data cleaning treatment on the historical daily average output of the wind power station by combining historical wind speed data by adopting a variable point method-quartile method, removing invalid information caused by communication faults, artificial wind abandon and data recording errors, and obtaining a wind speed-daily average output curve of the wind power station; for the day with abnormal wind power, linear interpolation is carried out by adopting a linear interpolation method to obtain daily average output, and finally, a wind speed-wind power time sequence data set W { (v) which can be used for training is obtained1,p1),(v2,p2),…,(vi,pi),…,(vN,pN) Where v is 1,2, …, NiWind speed at day i, piIs the wind power on day i;
step (3), p in the sequence W1,p2,…,pnPerforming sequence decomposition on part of the I eigen-mode functions by adopting an empirical mode decomposition method to obtain corresponding I eigen-mode functions and a remainder;
step (4), for the 2 nd eigenmode function to the I th eigenmode function and the residual expression, i.e. IMF2、……、IMFI、R0The corresponding mathematical function of which is expressed as f1(vi),f2(vi),…,fI(vi) Respectively forming T by the wind speed values in the sequence W1、……、TII new input training set Tm={vi,fm(vi)},m=1,2, …, I; determining a forecast period T for model predictionsforeSetting the number of neurons in the input layer of the training model corresponding to each eigenmode function as TforeThe number of neurons in the output layer is 1, and T is set for each training setiRespectively carrying out BP neural network training; finally, obtaining a wind speed-wind power eigenmode function BP neural network model suitable for power station wind power prediction;
step (5), predicting the day-ahead T of the starting point according to the input condition required in the model training prediction in the step (4)foreInputting the historical wind speed sequence of the day into the trained model (4), and predicting backward T from the prediction starting point by the model day by dayforeAnd (4) average output of the wind power station in each day, namely the predicted wind power generation capacity in the period to be predicted of the wind power station.
The specific method of the step (2) is as follows:
step (2.1), grouping the historical daily average output data of the wind power station and the corresponding wind speed data from small to large according to the wind speed interval of every 1m/S, and dividing the data into S1、S2、…、Sk、…SLL groups are provided, and the data length in each group is Dk;
Step (2.2), for each group Sk(k-1, 2, …, L) according to the formulaCalculate the variance q of each data pointtWherein p istThe average output force on the day of the t day,the mean of the average forces over all days in the group was followed by calculation of the rate of change of variance Δ for each data pointt=|qt-qt-1|;
Step (2.3), the variance change rate delta is processed by adopting a least square methodtIdentifying the change point, setting deltatObeying a two-segment linear model; by DeltatThe sum of the squares of the differences between the observed value and the theoretical value of (2) is used as an objective function, and the time or position of the point where the sum of the squares of the differences reaches the minimum value is used as an estimate of the position of the change point of the data series to obtain eachGroup SkThe data sequence of (1) is changed into a point sequence number, and partial abnormal values after the sequence number are discarded;
step (2.3), cleaning the numerical value points corresponding to the average output of the abnormal days in the wind speed interval after the abnormal interval is abandoned by adopting a quartile method; the daily average output interval [ P ] in the wind speed interval processed by the variable point methodmin,Pmax]Dividing the average into four parts, and recording the values at the positions of three division points as Q1、Q2、Q3(ii) a By a quarter of a bit distance IQR=Q3-Q1Determining an inner limit [ F ] of the daily average output value1,Fu]=[Q1-1.5IQR,Q3+1.5IQR](ii) a And determining the wind speed-daily average output data points outside the inner limit as daily average output abnormal data points for cleaning.
In the step (4), the training step of the BP neural network training is that ① takes T forward from the prediction starting point dayforeInputting the daily wind speed data into the input layer of each eigenmode function BP neural network, ② obtaining the predicted value p of the 1 st daily average output of the predicted section from the output layerf1Corresponding eigenmode function values; adding the eigenmode function values to obtain pf1Obtaining a corresponding wind speed value v from a wind speed-daily average output curve f1③ moving the prediction starting point by 1 day, predicting the 2 nd, … … th and Tth according to the steps of ① and ②foreAnd (4) predicting the value.
As shown in fig. 4, the wind power generation capacity prediction system coupling meteorological information and EMD technology is characterized by including:
the data acquisition module 101 is used for acquiring daily average output of the wind power station, installed capacity of the wind power station and daily wind speed time sequence data for more than 2 years;
the first processing module 102 is configured to perform cleaning processing on the acquired data by using a variable point method-quartile method, perform linear interpolation on a wind power abnormal day and a wind power abnormal day by using a linear interpolation method to obtain daily average output, and finally obtain a wind speed-wind power time sequence data set which can be used for training;
the second processing module 103 is configured to perform sequence decomposition on a wind power part in the wind speed-wind power time sequence data set by using an empirical mode decomposition method, and perform BP neural network training on an input training set formed by a 2 nd eigenmode function, an I th eigenmode function, a remainder expression and wind speeds in the wind speed-wind power time sequence data set to obtain a BP neural network model;
and the wind power generation capacity prediction module 104 is used for predicting the wind power generation capacity in the period to be predicted by adopting a BP neural network model.
In the embodiment of the invention, a data acquisition module 101 acquires daily average output of a wind power station, installed capacity of the wind power station and daily wind speed time sequence data for more than 2 years; the first processing module 102 performs cleaning processing on the acquired data by adopting a variable point method-quartile method, performs linear interpolation on the day with abnormal wind power and the day with abnormal wind power by adopting a linear interpolation method to obtain daily average output, and finally obtains a wind speed-wind power time sequence data set which can be used for training; the second processing module 103 performs sequence decomposition on the wind power part in the wind speed-wind power time sequence data set by using an empirical mode decomposition method, and performs BP neural network training on an input training set formed by the 2 nd eigenmode function, the I th eigenmode function, the remainder expression and the wind speed in the wind speed-wind power time sequence data set to obtain a BP neural network model; the wind power generation capacity prediction module 104 predicts the wind power generation capacity in the to-be-predicted time period by adopting a BP neural network model.
According to the wind power generation capacity prediction system coupling meteorological information and an EMD technology, the system considers the physical influence factors of wind power generation while performing statistical analysis and soft computing method modeling on historical time sequence data, the prediction precision is greatly improved compared with a method of singly using one of the historical time sequence data and the soft computing method, the prediction requirement on the wind power generation capacity in the process of making a power grid scheduling plan is better met, and the wind power generation capacity prediction system has important use and popularization values.
The system provided by the embodiment of the present invention is used for executing the above method embodiments, and for details of the process and the details, reference is made to the above embodiments, which are not described herein again.
Fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, and referring to fig. 5, the electronic device may include: a processor (processor)201, a communication Interface (communication Interface)202, a memory (memory)203 and a communication bus 204, wherein the processor 201, the communication Interface 202 and the memory 203 complete communication with each other through the communication bus 204. The processor 201 may call logic instructions in the memory 203 to perform the following method: extracting daily average output, installed capacity and daily wind speed time sequence data of the wind power station which are acquired for more than 2 years; cleaning the collected data by adopting a variable point method-quartile method, and performing linear interpolation on the wind power abnormal day and the wind power abnormal day by adopting a linear interpolation method to obtain daily average output so as to finally obtain a wind speed-wind power time sequence data set for training; performing sequence decomposition on a wind power part in the wind speed-wind power time sequence data set by adopting an empirical mode decomposition method, and performing BP neural network training on an input training set formed by a 2 nd eigenmode function, an I th eigenmode function, a remainder expression and wind speed in the wind speed-wind power time sequence data set to obtain a BP neural network model; and predicting the wind power generation capacity in the period to be predicted by adopting a BP neural network model.
In addition, the logic instructions in the memory 203 may be implemented in the form of software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented to, when executed by a processor, perform the wind power generation capability prediction method for coupling meteorological information and EMD technology provided in the foregoing embodiments, for example, the method includes: extracting daily average output, installed capacity and daily wind speed time sequence data of the wind power station which are acquired for more than 2 years; cleaning the collected data by adopting a variable point method-quartile method, and performing linear interpolation on the wind power abnormal day and the wind power abnormal day by adopting a linear interpolation method to obtain daily average output so as to finally obtain a wind speed-wind power time sequence data set for training; performing sequence decomposition on a wind power part in the wind speed-wind power time sequence data set by adopting an empirical mode decomposition method, and performing BP neural network training on an input training set formed by a 2 nd eigenmode function, an I th eigenmode function, a remainder expression and wind speed in the wind speed-wind power time sequence data set to obtain a BP neural network model; and predicting the wind power generation capacity in the period to be predicted by adopting a BP neural network model.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
The method for predicting the wind power generation capacity by coupling meteorological information and an EMD data processing technology comprises three parts, namely, cleaning of training set data by a variable point method-a quartile method, wavelet decomposition of the training set data by EMD, and prediction model training according to a decomposition eigenmode function series.
① training set data cleaning based on variable point method-quartile method
The variable point method refers to a point where a certain amount or some amount changes suddenly in a sequence, and the change is often indicative of the change of data quality. And (3) dividing the problem into a variable point problem of a position parameter and a scale parameter according to the main digital characteristic change of the mean value and the variance of the analyzed sample data. In the problem of wind power prediction modeling, the invention researches position parameters of data characteristic mutation in wind speed-daily average output. If accumulation type abnormal data exist in a wind speed interval, the characteristics of the data such as the change rate, the mean value, the variance or the change rate of the variance of the power sequence are all mutated. When the variance change rate is used as a variable point grouping basis, the cleaning effect is good while the data loss amount is relatively small, so that the variable point identification is carried out on the variance change rate of the daily average output of the wind power plant, the position where the power value obviously changes can be obtained from a statistical angle, and a data set of normal wind power in a certain wind speed interval is further obtained.
The quartile is a numerical value at the position of three dividing points for averagely dividing one sequenced data sample into four parts and is respectively marked as Q1、Q2、Q3. The quartile distance I can be obtained through the calculated quartileQR=Q3-Q1Finally, the inner limit [ F ] of the abnormal value in the data sample is determined1,Fu]=[Q1-1.5IQR,Q3+1.5IQR]. At an inner limit [ F1,Fu]All the other values are determined as abnormal values.
By combining the two common data cleaning methods, the abnormal numerical values of concentrated accumulation and dispersion caused by communication errors, instrument recording errors, artificial wind abandoning and the like in the wind speed-wind power data series can be effectively eliminated, and a relatively accurate data set is provided for subsequent modeling analysis.
② preprocessing of wind farm power sequence-empirical mode decomposition algorithm (EMD)
In the process of training a model specifically, because the output time sequence of wind power has strong randomness from the statistical viewpoint, the inertia of the whole data series along with the time change is poor, and if the whole sequence is directly used for neural network modeling prediction, high-dimensional data characteristics are difficult to capture, so that the model building precision is reduced. Therefore, when the wind power generation capacity is predicted by using the historical wind power output data, the existing data sequence needs to be processed to remove the noise of the data sequence as much as possible and show regularity. For the problems, the invention provides a solution for processing the model input condition by combining with EMD, highlighting the regularity of wind power change and improving the modeling precision.
EMD is the core content of the hilbert-yellow transform, which assumes that any complex signal is composed of simple eigen-state functions IMF (intrinsic Mode function), and that IMFs are independent of each other. IMF has 2 features as follows: the number of the extreme points (maximum value and minimum value) in the whole function is equal to the number of the zero-crossing points or has 1 difference at most; at any point, the mean value of the corresponding points of the upper envelope composed of local maxima and the lower envelope composed of local minima is zero.
The EMD algorithm of the daily average output time sequence of the wind power station comprises the following steps:
a. for the historical exertion time series p (t), all its maximum and minimum points are first found. And respectively connecting all the maximum value points and the minimum value points into a curve by a cubic spline interpolation method, wherein the curve is called an upper envelope line and a lower envelope line. Let m (t) be the mean of the upper and lower envelope, h (t) be the difference between the sequences p (t) and m (t), i.e.: h (t) ═ p (t) -m (t).
b. Taking h (t) as newRepeating the process in step a k times until hk(t) is IMF. Judgment hkThe criteria for whether (t) is IMF are:
if R iskLess than a predetermined threshold value, hk(t) can be judged as an IMF. Let the 1 st IMF be the IMF1(t)=hk(t)。
c. Obtaining the IMF1After (t), the IMF is subtracted from the original daily average force sequence1(t) obtaining the remainder r1(t), namely: r is1(t)=p(t)-IMF1(t)
Will r is1(t) repeating steps a and b as z (t) to obtain the 2 nd IMF, i.e. IMF2(t) if r2(t)=r1(t)-IMF2(t) adding r2(t) as z (t) and then repeating the above steps. To obtain all IMFs, the process is repeated until the remainder function r (t) is a monotonic function or less than a predetermined threshold. The final remainder function r (t) is called the remainder.
In the research process, EMD decomposition is carried out on historical wind power data sequences of a plurality of power stations, and the result shows that the sequence corresponding to the high-frequency eigenmode function often has certain negative influence on the finally carried out neural network modeling and certain disturbance on the prediction precision of the model, so that IMF is abandoned in the next model training process1The wind power decomposition sequence of (2) is not included in the training of a prediction model. The decomposition result curve of the power process of the wind power of a section of 30d of the bean grinding mountain wind power field is shown in figure 2 after EMD decomposition.
③ BP modeling and training are carried out by combining each sequence of wind speed-wind power IMF
The model is established by adopting a BP neural network method, the input condition is a wind speed forecasting numerical sequence of the forecasting section time, and the output sequence is an eigen state function sequence value corresponding to each of N IMF functions. On the basis of processing historical output data by empirical mode decomposition, wind speed and IMF (inertial measurement function) are processed1、IMF2、……、IMFnRespectively establishing NAnd predicting by the BP model, and finally outputting a sequence value for reforming to obtain a power predicted value sequence of the wind power plant.
The BP neural network is a multi-layer feedforward type network, and is generally called a BP network because adjustment of network parameters employs a Back Propagation (BP) learning algorithm. Mathematicians have demonstrated that a three-layer neural network with sigmoid nonlinear transfer functions can arbitrarily approximate any continuous function. A typical BP network is shown in fig. 2.
After the network structure is determined, the network can be trained using the sample set. The essence of the training is a learning and adjusting process, namely, learning and adjusting the weight and the threshold of the network, so that the network searches for an inherent invisible functional relationship between the input and the output of the sample set. The BP learning algorithm essentially converts a group of sample input and output problems into a nonlinear optimization problem (called an expression function), and adjusts network parameters by using the gradient of the expression function, namely, the modification of weights and thresholds is along the direction of the fastest descending expression function, namely the direction of negative gradient. The learning process comprises two processes of error forward propagation and error backward propagation. In the forward propagation process, input information is processed layer by layer from an input layer through an implicit layer and is transmitted to an output layer, and the state of each layer of neurons only affects the state of the next layer of neurons. If the output layer can not obtain the expected output, the method shifts to the reverse propagation and returns the error signal along the original path. The error is minimized by iteratively modifying the weights and thresholds of the neurons of each layer. In the BP learning algorithm, the weight and threshold value of each layer of neurons are adjusted according to the following formula:
xk+1=xk-akgk
wherein x iskIs the current weight and threshold matrix, gkIs the gradient of the current performance function, akIs the learning rate.
Assuming a three-layer BP network, input node xiImplicit layer node yjOutput node zlThe weight matrix of the input layer and the hidden layer is wjiThe weight matrix of the hidden layer and the output node is vljOutput layer and implicit layerThe layer threshold is b, and the node function output values of the input layer and the output layer are respectively f (net)j)、f(netl)。
When the expected value of the output node is tlThe BP learning algorithm is derived as follows:
the output of the hidden layer node:
output of the output layer node:
calculation error of output layer node:
the weight correction formula is as follows:
wji(k+1)=wji(k)+Δwji=wji(k)+η′′jxi
with layer node error implicit'jIn (1)Representing an output node zlError of (2)lBy weight vljTo node yjBack-propagation becomes an error for the hidden layer node.
The modified formula of the output layer and hidden layer threshold b is as follows:
bl(k+1)=bl(k)+ηl
bj(k+1)=bj(k)+η′′j
therefore, the method comprises the following steps:
an output layer:
f′(netl)=f(netl)(1-f(netl))=zl(1-zl)
hidden layer:
f′(netj)=f(netj)(1-f(netj))=yj(1-yj)
the calculation process of the final training weight of the intermediate hidden neuron and the output neuron of the BP neural network algorithm is described above.
Examples of the applications
According to the calculation method, the wind power generation output process of the future 10d of an actual wind power station is predicted and calculated. In the example, historical operating data and wind speed data of a Yunnan grinding bean mountain wind power plant are used. The predicted results are shown in table 1 below. In order to compare the rationality of the results obtained by the EMD-BP method, the prediction model calibration and 10d wind power prediction are carried out on the calculation example under the same condition by adopting a method of directly carrying out BP neural network training without carrying out EMD decomposition on the wind power series, and the results are shown in the column 4 of the table above. The comparison of the predicted course and the actual course of the force is shown in fig. 3. As can be seen from comparison of prediction results, for days 2.1 to 2.5, the average prediction accuracy of the EMD-BP method is 89.9%, while the prediction accuracy of the BP method is 89.84%, and the prediction accuracy is relatively close. On the 5 th day after the long prediction period, the average prediction accuracy of the EMD-BP method is 87.4%, and the average prediction accuracy of the BP method is 80.2%.
TABLE 1 Mozuki 10d prediction results Table
According to the actual prediction calculation, in the 5d close to the prediction starting point part, the prediction value of the prediction model is close to the actual value, the size and the change trend are consistent, and the prediction accuracy of the prediction model on the wind power output is relatively high. However, if the distance from the prediction starting point is long, the previous prediction value appears in the input condition, and the magnitude or the trend of the numerical value changes greatly, the prediction accuracy of the part is deficient. The method of combining EMD empirical mode decomposition and BP neural network adopted by the calibration of the prediction model belongs to an excavation algorithm of data high-dimensional characteristics, and has higher prediction precision for time series data which has better continuity and can correctly reflect the physical change rule of actual data. However, the model does not make any research into the physical cause of the mapping relationship between the input and the output, so that the wind power at a time point far away from the existing actual data cannot be accurately predicted, and the wind power prediction accuracy at a mutation point and a later period caused by various human factors or sudden environmental changes is insufficient.
The foregoing shows and describes the general principles, essential features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (6)
1. The wind power generation capacity prediction method based on the coupling of meteorological information and EMD technology is characterized by comprising the following steps:
step (1), basic data required by a model is established, wherein the basic data comprises daily average output of a wind power station for more than 2 years, installed capacity of the wind power station and daily wind speed time sequence data, and the length of the data sequence is set to be N days;
step (2), carrying out data cleaning treatment on the historical daily average output of the wind power station by combining historical wind speed data by adopting a variable point method-quartile method, removing invalid information caused by communication faults, artificial wind abandon and data recording errors, and obtaining a wind speed-daily average output curve of the wind power station; for the day with abnormal wind power, linear interpolation is carried out by adopting a linear interpolation method to obtain daily average output, and finally, a wind speed-wind power time sequence data set W { (v) which can be used for training is obtained1,p1),(v2,p2),…,(vi,pi),…,(vN,pN)}(i=1,2, …, N), wherein viWind speed at day i, piIs the wind power on day i;
step (3), p in the sequence W1,p2,…,pnPerforming sequence decomposition on part of the I eigen-mode functions by adopting an empirical mode decomposition method to obtain corresponding I eigen-mode functions and a remainder;
step (4), for the 2 nd eigenmode function to the I th eigenmode function and the residual expression, i.e. IMF2、……、IMFI、R0The corresponding mathematical function of which is expressed as f1(vi),f2(vi),…,fI(vi) Respectively forming T by the wind speed values in the sequence W1、……、TII new input training set Tm={vi,fm(vi) 1,2, …, I; determining a forecast period T for model predictionsforeSetting the number of neurons in the input layer of the training model corresponding to each eigenmode function as TforeThe number of neurons in the output layer is 1, and T is set for each training setiRespectively carrying out BP neural network training; finally, obtaining a wind speed-wind power eigenmode function BP neural network model suitable for power station wind power prediction;
step (5), predicting the day-ahead T of the starting point according to the input condition required in the model training prediction in the step (4)foreInputting the historical wind speed sequence of the day into the trained model (4), and predicting backward T from the prediction starting point by the model day by dayforeAnd (4) average output of the wind power station in each day, namely the predicted wind power generation capacity in the period to be predicted of the wind power station.
2. The method for predicting the wind power generation capacity by coupling meteorological information and EMD technology according to claim 1, wherein the specific method in the step (2) is as follows:
step (2.1), grouping the historical daily average output data of the wind power station and the corresponding wind speed data from small to large according to the wind speed interval of every 1m/S, and dividing the data into S1、S2、…、Sk、…SLL groups are provided, and the data length in each group is Dk;
Step (2.2), for each group Sk(k-1, 2, …, L) according to the formulaCalculate the variance q of each data pointtWherein p istThe average output force on the day of the t day,the mean of the average forces over all days in the group was followed by calculation of the rate of change of variance Δ for each data pointt=qt-qt-1|;
Step (2.3), the variance change rate delta is processed by adopting a least square methodtIdentifying the change point, setting deltatObeying a two-segment linear model; by DeltatThe sum of the squares of the differences between the observed value and the theoretical value of (A) is used as an objective function, and the time or position of the point where the sum of the squares of the differences reaches the minimum value is used as an estimate of the position of the change point of the data series, to determine the S for each groupkThe data sequence of (1) is changed into a point sequence number, and partial abnormal values after the sequence number are discarded;
step (2.3), cleaning the numerical value points corresponding to the average output of the abnormal days in the wind speed interval after the abnormal interval is abandoned by adopting a quartile method; the daily average output interval [ P ] in the wind speed interval processed by the variable point methodmin,Pmax]Dividing the average into four parts, and recording the values at the positions of three division points as Q1、Q2、Q3(ii) a By a quarter of a bit distance IQR=Q3-Q1Determining an inner limit [ F ] of the daily average output value1,Fu]=[Q1-1.5IQR,Q3+1.5IQR](ii) a And determining the wind speed-daily average output data points outside the inner limit as daily average output abnormal data points for cleaning.
3. The method for predicting wind power generation capacity by coupling meteorological information and EMD technology according to claim 1, wherein in the step (4), the training step of BP neural network training is that ① takes T forward from the prediction starting point dayforeInputting the daily wind speed data into the input layer of each eigenmode function BP neural network, ② obtaining the predicted value p of the 1 st daily average output of the predicted section from the output layerf1Corresponding eigenmode function values; adding the eigenmode function values to obtain pf1Obtaining a corresponding wind speed value v from a wind speed-daily average output curvef1③ moving the prediction starting point by 1 day, predicting the 2 nd, … … th and Tth according to the steps of ① and ②foreAnd (4) predicting the value.
4. Wind power generation capacity prediction system of coupling meteorological information and EMD technique, its characterized in that includes:
the data acquisition module is used for acquiring daily average output of the wind power station, installed capacity of the wind power station and daily wind speed time sequence data for more than 2 years;
the first processing module is used for cleaning the acquired data by adopting a variable point method-quartile method, and for the abnormal wind power day, performing linear interpolation by adopting a linear interpolation method to obtain daily average output, and finally obtaining a wind speed-wind power time sequence data set for training;
the second processing module is used for performing sequence decomposition on the wind power part in the wind speed-wind power time sequence data set by adopting an empirical mode decomposition method, and performing BP neural network training on an input training set formed by the 2 nd eigenmode function, the I th eigenmode function, the remainder expression and the wind speed in the wind speed-wind power time sequence data set to obtain a BP neural network model;
and the wind power generation capacity prediction module is used for predicting the wind power generation capacity in the period to be predicted by adopting a BP neural network model.
5. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements the steps of the wind power generation capability prediction method of any of claims 1 to 3 for coupling meteorological information with EMD techniques.
6. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the steps of the method for predicting wind power generation capability of coupling meteorological information and EMD techniques according to any one of claims 1 to 3.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010182883.8A CN111476402A (en) | 2020-03-16 | 2020-03-16 | Wind power generation capacity prediction method coupling meteorological information and EMD technology |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010182883.8A CN111476402A (en) | 2020-03-16 | 2020-03-16 | Wind power generation capacity prediction method coupling meteorological information and EMD technology |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111476402A true CN111476402A (en) | 2020-07-31 |
Family
ID=71747461
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010182883.8A Pending CN111476402A (en) | 2020-03-16 | 2020-03-16 | Wind power generation capacity prediction method coupling meteorological information and EMD technology |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111476402A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115545362A (en) * | 2022-12-05 | 2022-12-30 | 南方电网数字电网研究院有限公司 | AI and TSD combined new energy medium-term power combined prediction method |
CN117117858A (en) * | 2023-10-18 | 2023-11-24 | 中国华能集团清洁能源技术研究院有限公司 | Wind turbine generator power prediction method, device and storage medium |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103034757A (en) * | 2012-12-02 | 2013-04-10 | 中国科学院电工研究所 | Wind farm time-frequency domain modeling method based on empirical mode decomposition (EMD) |
CN103049798A (en) * | 2012-12-05 | 2013-04-17 | 浙江大学城市学院 | Short-period electric generation power forecasting method applied to photovoltaic electric generation system |
CN106295798A (en) * | 2016-08-29 | 2017-01-04 | 江苏省电力试验研究院有限公司 | Empirical mode decomposition and Elman neural network ensemble wind-powered electricity generation Forecasting Methodology |
CN108412710A (en) * | 2018-01-30 | 2018-08-17 | 同济大学 | A kind of Wind turbines wind power data cleaning method |
CN109919421A (en) * | 2019-01-23 | 2019-06-21 | 华南理工大学 | Short-term power load prediction model establishment method based on VMD-PSO-BPNN |
CN110059844A (en) * | 2019-02-01 | 2019-07-26 | 东华大学 | Energy storage device control method based on set empirical mode decomposition and LSTM |
CN110365053A (en) * | 2019-08-09 | 2019-10-22 | 南京信息工程大学 | Short-term wind power forecast method based on delay optimisation strategy |
CN110458316A (en) * | 2019-05-21 | 2019-11-15 | 武汉大学 | A kind of offshore wind farm power short term prediction method based on set empirical mode decomposition and LSTM network |
-
2020
- 2020-03-16 CN CN202010182883.8A patent/CN111476402A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103034757A (en) * | 2012-12-02 | 2013-04-10 | 中国科学院电工研究所 | Wind farm time-frequency domain modeling method based on empirical mode decomposition (EMD) |
CN103049798A (en) * | 2012-12-05 | 2013-04-17 | 浙江大学城市学院 | Short-period electric generation power forecasting method applied to photovoltaic electric generation system |
CN106295798A (en) * | 2016-08-29 | 2017-01-04 | 江苏省电力试验研究院有限公司 | Empirical mode decomposition and Elman neural network ensemble wind-powered electricity generation Forecasting Methodology |
CN108412710A (en) * | 2018-01-30 | 2018-08-17 | 同济大学 | A kind of Wind turbines wind power data cleaning method |
CN109919421A (en) * | 2019-01-23 | 2019-06-21 | 华南理工大学 | Short-term power load prediction model establishment method based on VMD-PSO-BPNN |
CN110059844A (en) * | 2019-02-01 | 2019-07-26 | 东华大学 | Energy storage device control method based on set empirical mode decomposition and LSTM |
CN110458316A (en) * | 2019-05-21 | 2019-11-15 | 武汉大学 | A kind of offshore wind farm power short term prediction method based on set empirical mode decomposition and LSTM network |
CN110365053A (en) * | 2019-08-09 | 2019-10-22 | 南京信息工程大学 | Short-term wind power forecast method based on delay optimisation strategy |
Non-Patent Citations (1)
Title |
---|
孙祥晟: "基于经验模态分解的神经网络光伏发电预测方法研究", 《电气技术》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115545362A (en) * | 2022-12-05 | 2022-12-30 | 南方电网数字电网研究院有限公司 | AI and TSD combined new energy medium-term power combined prediction method |
CN117117858A (en) * | 2023-10-18 | 2023-11-24 | 中国华能集团清洁能源技术研究院有限公司 | Wind turbine generator power prediction method, device and storage medium |
CN117117858B (en) * | 2023-10-18 | 2024-01-30 | 中国华能集团清洁能源技术研究院有限公司 | Wind turbine generator power prediction method, device and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109146162B (en) | A kind of probability wind speed forecasting method based on integrated Recognition with Recurrent Neural Network | |
CN110751318B (en) | Ultra-short-term power load prediction method based on IPSO-LSTM | |
CN110414045A (en) | Short-term wind speed forecasting method based on VMD-GRU | |
CN110826791A (en) | Hybrid wind power prediction method based on long-time and short-time memory neural network | |
CN111461463B (en) | Short-term load prediction method, system and equipment based on TCN-BP | |
CN111027775A (en) | Step hydropower station generating capacity prediction method based on long-term and short-term memory network | |
CN110717610B (en) | Wind power prediction method based on data mining | |
CN110942194A (en) | Wind power prediction error interval evaluation method based on TCN | |
CN111428201B (en) | Prediction method for time series data based on empirical mode decomposition and feedforward neural network | |
CN114757427A (en) | Autoregressive corrected LSTM intelligent wind power plant ultra-short term power prediction method | |
CN113344288B (en) | Cascade hydropower station group water level prediction method and device and computer readable storage medium | |
CN112580588A (en) | Intelligent flutter signal identification method based on empirical mode decomposition | |
CN111476402A (en) | Wind power generation capacity prediction method coupling meteorological information and EMD technology | |
CN110866633A (en) | Micro-grid ultra-short term load prediction method based on SVR support vector regression | |
CN111242353A (en) | Wind power combined prediction modeling and prediction method | |
CN115204035A (en) | Generator set operation parameter prediction method and device based on multi-scale time sequence data fusion model and storage medium | |
CN111506868B (en) | Ultra-short-term wind speed prediction method based on HHT weight optimization | |
CN113627685B (en) | Wind driven generator power prediction method considering wind power internet load limit | |
CN117114160A (en) | Short-term photovoltaic power prediction method | |
CN112116171A (en) | Novel photovoltaic power generation power prediction method based on neural network | |
CN116632820A (en) | Wind power prediction method based on sequence extension VMD and wind speed fluctuation characteristics | |
CN113762591A (en) | Short-term electric quantity prediction method and system based on GRU and multi-core SVM counterstudy | |
CN113128666A (en) | Mo-S-LSTMs model-based time series multi-step prediction method | |
CN109685242B (en) | Photovoltaic ultra-short term combined prediction method based on Adaboost algorithm | |
CN116646927A (en) | Wind power prediction method based on segmented filtering and longitudinal and transverse clustering |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200731 |
|
RJ01 | Rejection of invention patent application after publication |