CN108345996A - A kind of system and method reducing wind power checking energy - Google Patents
A kind of system and method reducing wind power checking energy Download PDFInfo
- Publication number
- CN108345996A CN108345996A CN201810118649.1A CN201810118649A CN108345996A CN 108345996 A CN108345996 A CN 108345996A CN 201810118649 A CN201810118649 A CN 201810118649A CN 108345996 A CN108345996 A CN 108345996A
- Authority
- CN
- China
- Prior art keywords
- wind power
- data
- wind
- analysis
- prediction
- 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.)
- Granted
Links
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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/80—Management or planning
- Y02P90/82—Energy audits or management systems therefor
-
- 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
Abstract
The present invention provides a kind of method reducing wind power checking energy, including：(1) wind power is carried out by the examination analysis of causes：It is analyzed including weather forecast, wind power prediction analysis, lexical analysis, data report analysis and the analysis of manual breakdown judge；(2) generating reduces the strategy of wind power examination：Being obtained according to the reason of analysis reduces the strategy of wind power checking energy.Also disclose the corresponding system for reducing wind power checking energy, wind-powered electricity generation investor loses caused by reducing examination of the grid company to wind farm power prediction precision, cost of verification and evaluation monthly is reduced using existence conditions to the examination of non-wind field reason proposition application of exempting from examination.
Description
Technical field
The invention belongs to electric system, more particularly to a kind of system and method reducing wind power checking energy.
Background technology
Accurate wind power prediction can help dispatching of power netwoks department to carry out the operation plan of all kinds of power supplys, improve power grid
The stability of operation, improves the ability of power grid consumption wind-powered electricity generation, and then reduces due to the economic damage brought to Wind Power Generation quotient of rationing the power supply
It loses, to increase wind power plant rate of return on investment, supplementary means is provided for the management work of wind power plant.
Wind power prediction can there are many mode classifications, are divided into first prediction of wind speed then basis according to the physical quantity of prediction
Wind turbines or wind power curve prediction output power, and directly predict output power；It is classified as continuing by mathematical model
Prediction technique, ARAM models, i.e. difference ARMA model are mostly used in Random time sequence method, Kalman filtering with
And intelligent method, such as artificial neural network；Be divided into according to input data do not use data data of weather forecast based on time sequence
Row and using data weather forecast physics, statistics and integrated approach；Temporally scale is divided into ultra-short term prediction and short-term
Prediction.
Wind-powered electricity generation forecasting system is examined by certain technological means, especially important, the core of wind-powered electricity generation prediction examination is asked
Topic is the comparison of the generation schedule and the production same day practical wind power output of examining wind power plant to report.The generation schedule that wind power plant reports
It is based on wind power checking system, what this system prediction obtained is a possible wind power curve, the production same day
Practical wind power output is indicated by the ground data of template processing machine plan in power plant.At present by with a varied topography, near-earth wind speed numerical value is pre-
Report that difficulty is big, numerical weather forecast precision is relatively low, Predicting Technique research starting is late, prediction technique to data quality requirement height etc. because
The restriction of element, wind power prediction is horizontal universal relatively low at present, forecasts to manage about wind farm power prediction according to National Energy Board
The requirement of Tentative Measures is managed, the root-mean-square error of whole day prediction result should be less than 20%.Flat Shandong Wolong hole wind power plant 2015 12
The moon puts into operation on the 20th, and in July, 2016 enters assessment period, multi-party by the maturity of wind power prediction systems technology and wind-resources etc.
The influence of face factor, every month, the wind power prediction of wind power plant was examined substantially all in by examination state, 7 monthly examination nuclear powers in 2016
Measure 28.19MWH, the checking energy of in August, 2016 129.16MWH, the checking energy of in September, 2016 13.91MWH, 10 monthly examinations in 2016
Nuclear power amount 6.35MWH, total checking energy 177.61MWH in 2016.Therefore, it is necessary to search wind power prediction rate of accurateness not
The reason of meeting power grid check requirements, grasps prediction rule, and adjustment reports prediction to contribute, and analyzes the specific data examined, right
The examination proposition application of exempting from examination of non-wind field reason reduces cost of verification and evaluation monthly using existence conditions.
Invention content
The technical problem to be solved by the present invention is to：A kind of reduction wind power is overcome the deficiencies of the prior art and provide to examine
The system and method for nuclear power amount, wind-powered electricity generation investor damages caused by reducing examination of the grid company to wind farm power prediction precision
It loses, cost of verification and evaluation monthly is reduced using existence conditions to the examination of non-wind field reason proposition application of exempting from examination.
For this purpose, the purpose of the present invention is to provide a kind of systems reducing wind power checking energy, including：
(1) analysis of causes system：Including weather forecast analysis system, wind power prediction analysis system, lexical analysis system
System, data report the analysis system and manual breakdown judge analysis system, the result of the analysis to include：(1) weather forecast is sent out
Send unsuccessful and data update not in time；(2) wind power forecasting system is not perfect, and wind-resources are unstable；(3) scheduling system is former
Cause, including peak regulation, dispatch data net interruption are participated in, not in time, upload rate is imperfect for data update；(4) reported data is accurate
Rate can not count before reporting；(5) personnel's judgement, troubleshooting capability are low；
(2) strategy generating system：Being obtained according to the result of analysis of causes network analysis reduces wind power checking energy
Strategy, the strategy include：(1) operator on duty pays close attention to, and carries out system program update and optimization, upgrade-system in time；
(2) optimize wind power forecasting system, adjust Forecast Model For Weather, according to the history measured data of the anemometer tower at scene, to more
The different meteorological sources of kind are preferentially chosen, and according to the actual power data of history, adjustment is optimized to power prediction model,
Ensure the generated output feature that the prediction data of onsite application can withdraw deposit recent；(3) of that month wind power examination is verified again
Data；(4) it checks the previous day data accuracy, carries out data and repay analysis；(5) failure problems Macro or mass analysis improves live people
Member's troubleshooting is horizontal, shortens fault time.
Preferably, include after the wind power forecasting system optimization：
(1) data acquisition server, including data collector are used for operation data acquisition software, with wind farm side wind-powered electricity generation
Driven integrated communication management terminal communicates and acquires wind turbine, anemometer tower, wind power, numerical weather forecast, wind power plant local wind-powered electricity generation
Power prediction result data；
(2) database server：It is described to ensure data reliable memory for the processing, statistical analysis and storage of data
Database server configures disk array；
(3) work station is applied：Including PC station devices, the modeling of system is completed, graphic hotsopt is shown, report making is beaten
Print function；
(4) wind power prediction server：Wind power prediction module is run using property server, based on acquiring or carry
For the numerical weather forecast that SCADA system provides, using based on Weighted Least Squares Support Vector Machines and quantum particle swarm prediction
Neural network integration algorithm, and combine the real time execution operating mode of current wind electric field blower to separate unit wind turbine and entire wind power plant
Output situation carries out short-term forecast and ultra-short term prediction；
(5) data interface server：Numerical weather forecast is obtained using property server；
(6) reversed physics isolation technology, including the network switch and network communication attachment, physical isolation apparatus, cabinet with
Attachment is arranged at network boundary for ensureing network security, at the same to data acquisition and supervisor control (SCADA) and
Energy Management System (EMS) transmits the result of wind power prediction.
Preferably, the wind power forecasting system after the optimization further includes anemometer tower and curve display module, described
Anemometer tower is for measuring Carry Meteorological data, and to carry out ultra-short term power prediction, the entity anemometer tower is according to the reality of wind field
Border physical condition installation is mounted within the scope of wind field 5km to ensure the accuracy of prediction, passes through GPRS or collecting fiber wind
The real time meteorological data of tower, anemometer tower height are not less than axial fan hub height, and the anemometer tower hardware support kit equipment includes meteorology
Data pick-up, data acquisition equipment, data transmission set, including 3 air velocity transducers, 2 wind transducers, 1 air
Temperature sensor, 1 atmospheric humidity sensor and 1 barometric pressure sensor and data collector and data transmission set, institute
It states atmosphere temperature transducer and barometric pressure sensor is mounted in 8 meters of height；10 meters of height fill a wind transducer, other installations
Height installs one or more wind transducers；Air velocity transducer installs one respectively at 10 meters and 30 meters, to the wind-powered electricity generation work(
The result of rate prediction is shown using curve display module, and the curve includes prediction and actual curve, in the form of a list
Each wind field current prediction power and actual power are illustrated, and date control, the drop-down list of wind speed floor height can be passed through
Etc. conditions carry out multiple choices, and be saved in and local carry out subsequent query.
The present invention also aims to provide a kind of method reducing wind power checking energy, including：
(1) wind power is carried out by the examination analysis of causes：It is analyzed including weather forecast, wind power prediction analysis, scheduling
Analysis, data report analysis and the analysis of manual breakdown judge；
(2) generating reduces the strategy of wind power examination：Being obtained according to the reason of analysis reduces wind power checking energy
Strategy.
Preferably, wind power includes by the reason of examination analysis：(1) weather forecast sends unsuccessful and data update
Not in time；(2) wind power forecasting system is not perfect, and wind-resources are unstable；(3) system reason is dispatched, including participates in peak regulation, scheduling
Data network interrupts, and not in time, upload rate is imperfect for data update；(4) accuracy rate of reported data can not count before reporting；
(5) personnel's judgement, troubleshooting capability are low.
Preferably, the strategy includes：(1) operator on duty pays close attention to, and carries out system program update and optimization in time,
Upgrade-system；(2) optimize wind power forecasting system, adjust Forecast Model For Weather, surveyed according to the history of the anemometer tower at scene
Data preferentially choose a variety of different meteorological sources, according to the actual power data of history, are carried out to power prediction model
It optimizes and revises, ensures the generated output feature that the prediction data of onsite application can withdraw deposit recent；(3) of that month wind-powered electricity generation is verified again
Power examination data；(4) it checks the previous day data accuracy, carries out data and repay analysis；(5) failure problems Macro or mass analysis carries
High Field Force's troubleshooting is horizontal, shortens fault time.
Preferably, the optimization wind power forecasting system includes adding in systems using based on neural network prediction combination
The integration algorithm for weighing least square method supporting vector machine and quanta particle swarm optimization, includes the following steps：
(1) Neural Network model predictive meteorological data is established；
(2) the amphineura network model of prediction wind power data is established according to gained weather prognosis data；
(3) it determines the relationship between reference point and the wind speed time graph of wind power plant, predicts wind-powered electricity generation work(in the following a few hours
Rate.
Preferably, the step (1) includes：Historical data is pre-processed, including to missing data and " dirty " data
Processing, then meteorological condition is quantified, finally by after their normalizeds, the feature vector for constituting one day is input to
Amphineura network prediction model is trained and predicts.
Preferably, wherein the step (2) includes：
(2-1) encodes hyper parameter, including regularization parameter and nuclear parameter, each particle replaces with potential solution, forms one
A hyper parameter group, and select optimized parameter；
(2-2) establishes fitness function, assesses Generalization Capability；
After (2-3) iteration, the training of amphineura network model is carried out, training RNN is similar with common neural network,
Using back-propagation algorithm, because the parameter at each moment is shared, therefore the gradient of parameter does not only rely on current time
Output, also rely on before at the time of；
The extension of (2-4) amphineura network model：The thought of two-way extension is before the output of t moment not only depends on
Element, and the element after also relying on, output rely on the hidden state of two neural network models, therefore to amphineura network mould
Type carries out two-way extension.
Preferably, the step (2-2) fitness function is：Fitness=1/RMSE (γ, σ), wherein RMSE (γ,
σ) be prediction result root-mean-square error, with least square method supporting vector machine parameter to (γ, σ) change and change, work as termination
When Criterion of Iterative meets, maximum adaptation function just corresponds to least square method supporting vector machine optimized parameter, the termination of the algorithm
Criterion of Iterative includes two ways：First is that algorithm stops when target function value is less than or equal to a given threshold value ε；The
Two kinds are given iterations in advance, stop iteration when reaching this numerical value.
According to the following detailed description of specific embodiments of the present invention in conjunction with the accompanying drawings, those skilled in the art will be brighter
The above and other objects, advantages and features of the present invention.
Description of the drawings
Some specific embodiments that the invention will be described in detail by way of example and not limitation with reference to the accompanying drawings hereinafter.
Identical reference numeral denotes same or similar component or part in attached drawing.It should be appreciated by those skilled in the art that these
What attached drawing was not necessarily drawn to scale.The target and feature of the present invention will be apparent from view of following description taken together with the accompanying drawings,
In attached drawing：
Attached drawing 1 is the system block diagram according to the reduction wind power checking energy of the embodiment of the present invention；
Attached drawing 2 is according to the wind power forecasting system block diagram after the optimization of the embodiment of the present invention；
Attached drawing 3 is the method flow diagram according to the reduction wind power checking energy of the embodiment of the present invention；
Attached drawing 4 is according to the wind power forecasting method flow chart after the optimization of the embodiment of the present invention；
Attached drawing 5 is short-term and ultra-short term predicts flow chart according to the wind power after the optimization of the embodiment of the present invention.
Specific implementation mode
Attached drawing 1 is the system block diagram according to a kind of reduction wind power checking energy of the embodiment of the present invention, including：
(1) analysis of causes system：Including weather forecast analysis system, wind power prediction analysis system, lexical analysis system
System, data report the analysis system and manual breakdown judge analysis system, the result of the analysis to include：(1) weather forecast is sent out
Send unsuccessful and data update not in time；(2) wind power forecasting system is not perfect, and wind-resources are unstable；(3) scheduling system is former
Cause, including peak regulation, dispatch data net interruption are participated in, not in time, upload rate is imperfect for data update；(4) reported data is accurate
Rate can not count before reporting；(5) personnel's judgement, troubleshooting capability are low；
(2) strategy generating system：Being obtained according to the result of analysis of causes network analysis reduces wind power checking energy
Strategy, the strategy include：(1) operator on duty pays close attention to, and carries out system program update and optimization, upgrade-system in time；
(2) optimize wind power forecasting system, adjust Forecast Model For Weather, according to the history measured data of the anemometer tower at scene, to more
The different meteorological sources of kind are preferentially chosen, and according to the actual power data of history, adjustment is optimized to power prediction model,
Ensure the generated output feature that the prediction data of onsite application can withdraw deposit recent；(3) of that month wind power examination is verified again
Data；(4) it checks the previous day data accuracy, carries out data and repay analysis；(5) failure problems Macro or mass analysis improves live people
Member's troubleshooting is horizontal, shortens fault time.
Attached drawing 2 be according to the wind power forecasting system after a kind of optimization of the embodiment of the present invention, including：(1) data are adopted
Collect server, including data collector, is used for operation data acquisition software, it is logical with wind farm side wind power-driven integrated communication management terminal
Believe and acquire wind turbine, anemometer tower, wind power, numerical weather forecast, wind power plant local wind power prediction result data；
(2) database server：For the processing, statistical analysis and storage of data, to ensure data reliable memory, the database clothes
Business device configures disk array；(3) work station is applied：Including PC station devices, complete the modeling of system, graphic hotsopt is shown,
Report making printing function；(4) wind power prediction server：Wind power prediction module, base are run using property server
In the numerical weather forecast for acquiring or providing SCADA system offer, using based on Weighted Least Squares Support Vector Machines and quantum
The neural network integration algorithm of population prediction, and combine the real time execution operating mode of current wind electric field blower to separate unit wind turbine and whole
The output situation of a wind power plant carries out short-term forecast and ultra-short term prediction；(5) data interface server：It is obtained using property server
Value weather forecast；(6) reversed physics isolation technology, including the network switch and network communication attachment, physical isolation apparatus,
Cabinet and attachment are arranged at network boundary for ensureing network security, while to data acquisition and supervisor control
(SCADA) and Energy Management System (EMS) transmission wind power prediction result；(7) entity anemometer tower implements gas for measuring
Image data, to carry out ultra-short term power prediction, entity anemometer tower is installed according to the actual physics condition of wind field, to ensure prediction
Accuracy, be mounted within the scope of wind field 5km, pass through GPRS or the real time meteorological data of collecting fiber wind tower, anemometer tower is high
Degree is not less than axial fan hub height.Anemometer tower hardware support kit equipment includes meteorological data sensor, data acquisition equipment, data biography
Transfer device, including 3 air velocity transducers, 2 wind transducers, 1 atmosphere temperature transducer, 1 atmospheric humidity sensor and 1
Platform barometric pressure sensor and data collector and data transmission set, the atmosphere temperature transducer and barometric pressure sensor
Mounted in 8 meters of height；10 meters of height fill a wind transducer, the one or more wind transducers of other mounting heights installation；Wind
Fast sensor installs one respectively at 10 meters and 30 meters；(8) curve display module uses the result of the wind power prediction
Curve display module is shown, and the curve includes prediction and actual curve, illustrates each wind field in the form of a list and works as
Preceding prediction power and actual power, and a variety of choosings can be carried out by the conditions such as drop-down list of date control, wind speed floor height
It selects, and is saved in local progress subsequent query.
Attached drawing 3 is the flow chart for the method for reducing wind power checking energy, including：(1) wind power is carried out to be examined
The analysis of causes：It is analyzed including weather forecast, wind power prediction analysis, lexical analysis, data report analysis and manual failure
Discriminatory analysis；(2) generating reduces the strategy of wind power examination：Being obtained according to the reason of analysis reduces wind power checking energy
Strategy.
Attached drawing 4 is the flow chart according to the optimization wind power forecasting method of the embodiment of the present invention.Exactly existing wind-powered electricity generation work(
Rate forecasting system there are the problem of, require to look up the reason of wind power forecasting system accuracy rate is unsatisfactory for power grid check requirements,
Prediction rule is grasped, adjustment reports prediction to contribute, and analyzes the specific data examined, and exempts from examination to the examination proposition of non-wind field reason
Application, using existence conditions, reduces cost of verification and evaluation monthly, to propose the prediction technique of the wind power.
The mathematical concept and physical principle that the wind power forecasting method is related to are explained in detail as follows.Least square branch
It is statistical important achievement to hold vector machine model, and the training process of least square method supporting vector machine follows structural risk minimization
Vector machine inequality constraints is changed to equality constraint by principle, and empiric risk is changed to quadratic power by the first power of deviation, will be solved
Quadratic programming problem is converted into solution system of linear equations, avoids insensitive loss function, greatly reduces computation complexity, and
Arithmetic speed is higher than general support vector machines, and to greatly facilitate the solution of Lagrange multipliers α, former problem is that QP is asked
Topic, and the problem of be then a solution system of linear equations in LSSVM.Least square method supporting vector machine algorithm description is：For giving
Fixed training set,Linear equation in feature space is defined as：yi=wTφ
(xi)+bi, i=1,2 ... l；Regression problem can be expressed as：s.t.yi=ωT
φ(xi)+b+ξi, i=1,2 ... l, wherein ξi∈ R are error, and C ＞ 0 are penalty coefficient, play the role of adjusting error.
In wind power system prediction, which has to the data near operating point larger related at work
Property, and be not very big with the correlation far from operating point regions data.With the variation of system, the acquisition of new data so that with
The model that preceding off-line data sample is established can not accurate description system actual state, in order to enable model to accurately reflect
The current state of wind power system should constantly utilize obtained new data to establish the new model that can reflect system the present situation.
Least square method supporting vector machine but has lost robustness while improving standard supporting vector machine model, is
Treat different wind-powered electricity generation time-variable data with a certain discrimination, least square method supporting vector machine is weighted for consideration, the tool of operation
Body step is：(1) training data set { x is givenk,yk, k=1,2...N find out optimized parameter, and e is calculated for optimized parameterk
=αk/γ；(2) according to error ekDistribution situation calculate its Robust Estimation value(3) by ekWithDetermine corresponding weights vk；
(4) a is solved*And b*, provide final Nonlinear Prediction Models：
In quanta particle swarm optimization, particle has quantum behavior, and search capability is much better than traditional particle cluster algorithm.
In order to ensure that convergence, each particle must converge on respective p points, p=(p in quanta particle swarm optimization1,
p2,...pd), pdIt is the value that the particle is tieed up in d.Wherein：
The position of particle is found by following equation：In formula, u be distribution between zero and one one with
Machine number.A global point mbest is introduced to calculate the variable that the following iteration of particle walks, it defines the part of all ions most
The average value of good position：And mbest=(mbest1,mbest2,...,
mbestD),β is the shrinkage expansion factor in formula, and adjusting it can be with control convergence speed, β=0.5+
0.5(tmax-t)/tmax, tmaxIt is the maximum times of iteration, M is the size of particle group, and the position of final particle can be write as：
Referring to Fig. 5, above-mentioned Weighted Least Squares Support Vector Machines and quantum particle swarm are combined based on neural network prediction
Algorithm includes：(1) Neural Network model predictive meteorological data is established；(2) prediction wind-powered electricity generation is built again according to gained weather prognosis data
The amphineura network model of power data；(3) relationship between reference point and the wind speed time graph of wind power plant is determined, prediction is not
Carry out wind power in a few hours.Amphineura network mould wherein under Weighted Least Squares Support Vector Machines and quanta particle swarm optimization
Type foundation includes three key factors：(1) how hyper parameter is replaced with to the position of particle, that is, how to be encoded, how is (2)
Fitness function is defined come the advantages of assessing particle；(3) amphineura network model foundation how is carried out.Including：
(2-1) encodes hyper parameter, including regularization parameter and nuclear parameter, each particle replaces with potential solution, forms one
A hyper parameter group, and select optimized parameter；
(2-2) establishes fitness function, assesses Generalization Capability, and the fitness function that the present embodiment uses is：
Fitness=1/RMSE (γ, σ), wherein RMSE (γ, σ) are the root-mean-square errors of prediction result, it is with most
Small two, which multiply support vector machines parameter, changes (γ, σ) and changes.When termination Criterion of Iterative meets, maximum adaptation function is just right
Answer least square method supporting vector machine optimized parameter.
The termination Criterion of Iterative of algorithm includes two ways：First is given when target function value is less than or equal to one
Algorithm stops when threshold epsilon；It is for second given iterations in advance, stops iteration when reaching this numerical value.
After (2-3) iteration, the training of amphineura network model is carried out, training RNN is similar with common neural network,
Using back-propagation algorithm, because the parameter at each moment is shared, therefore the gradient of parameter does not only rely on current time
Output, also rely on before at the time of.Such as in order to calculate the gradient of wind power plant t=4 when prediction, it would be desirable to it is wrong forward
Transmit 3 moment.
The extension of (2-4) amphineura network model：The thought of two-way extension is before the output of t moment not only depends on
Element, and the element after also relying on, output rely on the hidden state of two neural network models, therefore to amphineura network mould
Type carries out two-way extension.
The background that the present embodiment proposes is based on the flat Shandong Wolong hole wind power plant to put into operation on December 20th, 2015, in July, 2016
Part enters assessment period, is influenced by many factors such as the wind power forecasting system maturity of technology and wind-resources, each
The wind power prediction examination of month wind power plant substantially all in by examination state, in July, 2016 checking energy 28.19MWH, 2016
Year August checking energy 129.16MWH, the checking energy of in September, 2016 13.91MWH, in October, 2016 checking energy 6.35MWH,
Total checking energy 177.61MWH in 2016.Therefore it is built using the load sample in wind power plant spring in 2017 and meteorological data
Vertical prediction model.The sample frequency that wind power data are provided from SCADA system is 12 seconds every point datas, using 2 institute of attached drawing
The wind power forecasting system shown carries out the point prediction of wind power plant wind power.Historical data is pre-processed, including to lacking
The processing for losing data and " dirty " data, then by meteorological condition (such as rainfall, maximum temperature, the minimum temperature, temperature) amount of progress
Change, the feature vector for after their normalizeds, constituting one day is finally input to amphineura network prediction model and is trained
And prediction, normalization formula are：
Wherein, L indicates the amount after normalization, LiIndicate actual value, Lmax, LminPoint
It Biao Shi not practical maximum and minimum value；[a, b] expression is normalized to section.
By common prediction error, i.e., root-mean-square error, mean absolute error and mean absolute percentage error are to pre-
It surveys result to compare and analyze, following 0.25 hour, 1 hour and 2 hours wind power is predicted, root-mean-square error point
Not Wei 2.410,5.481 and 7.807, mean absolute error is respectively 1.018,3.814 and 4.936；Average absolute percentage misses
Difference is respectively 15.17,26,35 and 28.17.
Thus it proves, which greatly improves, and the root-mean-square error of prediction reduces, and average absolute is missed
Difference reduces, and average relative error reduces, and applies in large-scale wind field super short-period wind power prediction, the comparison of prediction result
Analytical proof this method is effective, and system prediction precision improves, and reduces examination of the grid company to wind farm power prediction precision
Caused wind-powered electricity generation investor loss reduces monthly the examination proposition of non-wind field reason application of exempting from examination using existence conditions
Cost of verification and evaluation.
It, will not be by these embodiments although the present invention is described by reference to specific illustrative embodiment
Restriction and only limited by accessory claim.It should be understood by those skilled in the art that can be without departing from the present invention's
The embodiment of the present invention can be modified and be changed in the case of protection domain and spirit.
Claims (10)
1. a kind of system reducing wind power checking energy, it is characterised in that including：
(1) analysis of causes system：Including weather forecast analysis system, wind power prediction analysis system, lexical analysis system, number
Include according to analysis system and manual breakdown judge analysis system, the result of the analysis is reported：(1) weather forecast can not send
Work(and data update are not in time；(2) wind power forecasting system is not perfect, and wind-resources are unstable；(3) system reason, packet are dispatched
It includes and participates in peak regulation, dispatch data net interruption, not in time, upload rate is imperfect for data update；(4) accuracy rate of reported data can not
It is counted before reporting；(5) personnel's judgement, troubleshooting capability are low；
(2) strategy generating system：Being obtained according to the result of analysis of causes network analysis reduces the strategy of wind power checking energy,
The strategy includes：(1) operator on duty pays close attention to, and carries out system program update and optimization, upgrade-system in time；(2) excellent
Change wind power forecasting system, adjust Forecast Model For Weather, according to the history measured data of the anemometer tower at scene, to a variety of differences
Meteorological sources preferentially chosen, according to the actual power data of history, adjustment is optimized to power prediction model, is ensured existing
The generated output feature that the prediction data that field uses can withdraw deposit recent；(3) of that month wind power examination data is verified again；
(4) it checks the previous day data accuracy, carries out data and repay analysis；(5) failure problems Macro or mass analysis improves Field Force's troubleshooting
Level shortens fault time.
2. a kind of system reducing wind power checking energy according to claim 1, it is characterised in that the wind-powered electricity generation work(
Include after the optimization of rate forecasting system：
(1) data acquisition server, including data collector are used for operation data acquisition software, are integrated with wind farm side wind-powered electricity generation
Telecommunication management terminal communicates and acquires wind turbine, anemometer tower, wind power, numerical weather forecast, wind power plant local wind power
Prediction result data；
(2) database server：For the processing, statistical analysis and storage of data, to ensure data reliable memory, the data
Library server configures disk array；
(3) work station is applied：Including PC station devices, the modeling of system is completed, graphic hotsopt is shown, report making printing work(
Energy；
(4) wind power prediction server：Wind power prediction module is run using property server, based on acquisition or is provided
The numerical weather forecast that SCADA system provides uses what is predicted based on Weighted Least Squares Support Vector Machines and quantum particle swarm
Neural network integration algorithm, and the real time execution operating mode of current wind electric field blower is combined to go out separate unit wind turbine and entire wind power plant
Power situation carries out short-term forecast and ultra-short term prediction；
(5) data interface server：Numerical weather forecast is obtained using property server；
(6) reversed physics isolation technology, including the network switch and network communication attachment, physical isolation apparatus, cabinet and attachment,
For ensureing network security, it is arranged at network boundary, while being acquired and supervisor control (SCADA) and energy pipe to data
Reason system (EMS) transmits the result of wind power prediction.
3. a kind of system reducing wind power checking energy according to claim 2, it is characterised in that after the optimization
Wind power forecasting system further include anemometer tower and curve display module, the anemometer tower is for measuring Carry Meteorological number
According to carry out ultra-short term power prediction, the entity anemometer tower is installed according to the actual physics condition of wind field, to ensure prediction
Accuracy, be mounted within the scope of wind field 5km, pass through GPRS or the real time meteorological data of collecting fiber wind tower, anemometer tower is high
Degree is not less than axial fan hub height, and the anemometer tower hardware support kit equipment includes meteorological data sensor, data acquisition equipment, number
According to transmission device, including 3 air velocity transducers, 2 wind transducers, 1 atmosphere temperature transducer, 1 atmospheric humidity sensing
Device and 1 barometric pressure sensor and data collector and data transmission set, the atmosphere temperature transducer and atmospheric pressure
Sensor is mounted in 8 meters of height；10 meters of height fill a wind transducer, the one or more wind direction sensings of other mounting heights installation
Device；Air velocity transducer installs one respectively at 10 meters and 30 meters, and mould is shown using curve to the result of the wind power prediction
Block is shown, and the curve includes prediction and actual curve, illustrates the current pre- measurement of power of each wind field in the form of a list
Rate and actual power, and multiple choices can be carried out by the conditions such as drop-down list of date control, wind speed floor height, and be saved in
It is local to carry out subsequent query.
4. a kind of method reducing wind power checking energy is applied to according to any reduction wind-powered electricity generations of claim 1-3
The system of power checking energy, it is characterised in that including：
(1) wind power is carried out by the examination analysis of causes：It is analyzed including weather forecast, wind power prediction analysis, lexical analysis,
Data report analysis and the analysis of manual breakdown judge；
(2) generating reduces the strategy of wind power examination：Being obtained according to the reason of analysis reduces the plan of wind power checking energy
Slightly.
5. a kind of method reducing wind power checking energy according to claim 4, it is characterised in that wind power quilt
Examination analysis the reason of include：(1) weather forecast sends unsuccessful and data update not in time；(2) wind power prediction system
It unites not perfect, wind-resources are unstable；(3) system reason is dispatched, including participates in peak regulation, dispatch data net interruption, data update is too late
When, upload rate is imperfect；(4) accuracy rate of reported data can not count before reporting；(5) personnel's judgement, troubleshooting capability
It is low.
6. a kind of method reducing wind power checking energy according to claim 4, it is characterised in that the strategy
Including：(1) operator on duty pays close attention to, and carries out system program update and optimization, upgrade-system in time；(2) optimize wind power
Forecasting system, adjust Forecast Model For Weather, according to scene anemometer tower history measured data, to a variety of different meteorological sources into
Row is preferentially chosen, and according to the actual power data of history, optimizes adjustment to power prediction model, ensures the pre- of onsite application
The generated output feature that measured data can withdraw deposit recent；(3) of that month wind power examination data is verified again；(4) it checks previous
Day data accuracy rate carries out data and repays analysis；(5) failure problems Macro or mass analysis improves Field Force's troubleshooting level, shortens event
Downtime.
7. according to a kind of method of any reduction wind power checking energies of claim 4-6, it is characterised in that described
It includes in systems using based on neural network prediction combination weighted least-squares supporting vector to optimize wind power forecasting system
The integration algorithm of machine and quanta particle swarm optimization, includes the following steps：
(1) Neural Network model predictive meteorological data is established；
(2) the amphineura network model of prediction wind power data is established according to gained weather prognosis data；
(3) it determines the relationship between reference point and the wind speed time graph of wind power plant, predicts wind power in the following a few hours.
8. a kind of method reducing wind power checking energy according to claim 7, it is characterised in that the step (1)
Including：Historical data is pre-processed, includes the processing to missing data and " dirty " data, then by the meteorological condition amount of progress
Change, the feature vector for after their normalizeds, constituting one day is finally input to amphineura network prediction model and is trained
And prediction.
9. a kind of method reducing wind power checking energy according to claim 7, it is characterised in that the wherein described step
Suddenly (2) include：
(2-1) encodes hyper parameter, including regularization parameter and nuclear parameter, each particle replaces with potential solution, forms one and surpasses
Parameter group, and select optimized parameter；
(2-2) establishes fitness function, assesses Generalization Capability；
After (2-3) iteration, the training of amphineura network model is carried out, training RNN is similar with common neural network, uses
Back-propagation algorithm, because the parameter at each moment is shared, therefore the gradient of parameter does not only rely on the output at current time,
At the time of before also relying on；
The extension of (2-4) amphineura network model：The thought of two-way extension is the member before the output of t moment not only depends on
Element, and the element after also relying on, output rely on the hidden state of two neural network models, therefore to amphineura network model
Carry out two-way extension.
10. a kind of method reducing wind power checking energy according to claim 9, it is characterised in that the step
(2-2) described fitness function is：Fitness=1/RMSE (γ, σ), wherein RMSE (γ, σ) are that the root mean square of prediction result misses
Difference changes as least square method supporting vector machine parameter changes (γ, σ), when termination Criterion of Iterative meets, maximum adaptation
Function just corresponds to least square method supporting vector machine optimized parameter, and the termination Criterion of Iterative of the algorithm includes two ways：The
First, algorithm stops when target function value is less than or equal to a given threshold value ε；It is for second a given iteration in advance
Number stops iteration when reaching this numerical value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810118649.1A CN108345996B (en) | 2018-02-06 | 2018-02-06 | System and method for reducing wind power assessment electric quantity |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810118649.1A CN108345996B (en) | 2018-02-06 | 2018-02-06 | System and method for reducing wind power assessment electric quantity |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108345996A true CN108345996A (en) | 2018-07-31 |
| CN108345996B CN108345996B (en) | 2021-07-20 |
Family
ID=62960001
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810118649.1A Active CN108345996B (en) | 2018-02-06 | 2018-02-06 | System and method for reducing wind power assessment electric quantity |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108345996B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109242223A (en) * | 2018-11-26 | 2019-01-18 | 武汉理工光科股份有限公司 | The quantum support vector machines of city Public Buildings Fire Risk is assessed and prediction technique |
| CN111079980A (en) * | 2019-11-22 | 2020-04-28 | 天合云能源互联网技术(杭州)有限公司 | Optical power prediction method based on self-adaptive classification strategy and hybrid optimization SVR |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080228553A1 (en) * | 2007-03-12 | 2008-09-18 | Airtricity Holdings Limited | Method And System For Determination Of An Appropriate Strategy For Supply Of Renewal Energy Onto A Power Grid |
| CN102736593A (en) * | 2012-06-05 | 2012-10-17 | 吴光军 | Integrated platform system for remote management and control of wind power field cluster |
| CN102945508A (en) * | 2012-10-15 | 2013-02-27 | 风脉（武汉）可再生能源技术有限责任公司 | Model correction based wind power forecasting system and method |
| CN103617454A (en) * | 2013-11-21 | 2014-03-05 | 中能电力科技开发有限公司 | Wind power plant power forecast method according to numerical weather forecasts |
| CN105279582A (en) * | 2015-11-20 | 2016-01-27 | 中国水利水电第十四工程局有限公司 | An ultra-short-term wind electricity power prediction method based on dynamic correlation characteristics |
| CN105552970A (en) * | 2016-02-25 | 2016-05-04 | 华北电力科学研究院有限责任公司 | Method and apparatus for improving accuracy of predicting power of wind power station |
| US20160333854A1 (en) * | 2015-05-15 | 2016-11-17 | General Electric Company | Digital Twin Interface for Operating Wind Farms |
| CN106157165A (en) * | 2015-04-09 | 2016-11-23 | 华电电力科学研究院 | A kind of based on laser radar anemometer Wind turbines power curve examination appraisal procedure |
| CN106326529A (en) * | 2016-08-09 | 2017-01-11 | 广东电网有限责任公司电力科学研究院 | System and method for assessing contribution electric quantity of primary frequency modulation of hydroelectric generating set |
| CN107480833A (en) * | 2017-09-05 | 2017-12-15 | 清华大学 | A kind of wind-powered electricity generation electricity generation system peak modulation capacity appraisal procedure |
-
2018
- 2018-02-06 CN CN201810118649.1A patent/CN108345996B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080228553A1 (en) * | 2007-03-12 | 2008-09-18 | Airtricity Holdings Limited | Method And System For Determination Of An Appropriate Strategy For Supply Of Renewal Energy Onto A Power Grid |
| CN102736593A (en) * | 2012-06-05 | 2012-10-17 | 吴光军 | Integrated platform system for remote management and control of wind power field cluster |
| CN102945508A (en) * | 2012-10-15 | 2013-02-27 | 风脉（武汉）可再生能源技术有限责任公司 | Model correction based wind power forecasting system and method |
| CN103617454A (en) * | 2013-11-21 | 2014-03-05 | 中能电力科技开发有限公司 | Wind power plant power forecast method according to numerical weather forecasts |
| CN106157165A (en) * | 2015-04-09 | 2016-11-23 | 华电电力科学研究院 | A kind of based on laser radar anemometer Wind turbines power curve examination appraisal procedure |
| US20160333854A1 (en) * | 2015-05-15 | 2016-11-17 | General Electric Company | Digital Twin Interface for Operating Wind Farms |
| CN105279582A (en) * | 2015-11-20 | 2016-01-27 | 中国水利水电第十四工程局有限公司 | An ultra-short-term wind electricity power prediction method based on dynamic correlation characteristics |
| CN105552970A (en) * | 2016-02-25 | 2016-05-04 | 华北电力科学研究院有限责任公司 | Method and apparatus for improving accuracy of predicting power of wind power station |
| CN106326529A (en) * | 2016-08-09 | 2017-01-11 | 广东电网有限责任公司电力科学研究院 | System and method for assessing contribution electric quantity of primary frequency modulation of hydroelectric generating set |
| CN107480833A (en) * | 2017-09-05 | 2017-12-15 | 清华大学 | A kind of wind-powered electricity generation electricity generation system peak modulation capacity appraisal procedure |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109242223A (en) * | 2018-11-26 | 2019-01-18 | 武汉理工光科股份有限公司 | The quantum support vector machines of city Public Buildings Fire Risk is assessed and prediction technique |
| CN111079980A (en) * | 2019-11-22 | 2020-04-28 | 天合云能源互联网技术(杭州)有限公司 | Optical power prediction method based on self-adaptive classification strategy and hybrid optimization SVR |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108345996B (en) | 2021-07-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108388962B (en) | Wind power prediction system and method | |
| Dong et al. | Wind power day-ahead prediction with cluster analysis of NWP | |
| CN108593990B (en) | Electricity stealing detection method based on electricity consumption behavior mode of electric energy user and application | |
| CN102945508B (en) | Model correction based wind power forecasting method | |
| CN103810533B (en) | A kind of distribution network failure Risk Identification Method based on cloud model | |
| CN107194625B (en) | Wind power plant wind curtailment electric quantity evaluation method based on neural network | |
| CN106779129A (en) | A kind of Short-Term Load Forecasting Method for considering meteorologic factor | |
| CN113496311A (en) | Photovoltaic power station generated power prediction method and system | |
| CN102509027B (en) | Wind powder combined predication method based on cross entropy theory | |
| CN112926770A (en) | Unified data metering and collecting system based on digital twins | |
| CN111985701A (en) | Power utilization prediction method based on power supply enterprise big data model base | |
| CN114118571A (en) | Heating heat load prediction method based on LSTM deep learning | |
| CN105184388A (en) | Non-linear regression method for urban power load short-period prediction | |
| CN105427005A (en) | Operation risk assessment method of wind power station | |
| CN110070223A (en) | A kind of short term power prediction technique applied to newly-built wind power plant | |
| CN110210679A (en) | A kind of load prediction system applied to County Level Distribution Network planning | |
| CN115481918A (en) | Active sensing and predictive analysis system for unit state based on source network load storage | |
| CN115545333A (en) | Method for predicting load curve of multi-load daily-type power distribution network | |
| CN108345996A (en) | A kind of system and method reducing wind power checking energy | |
| CN114021848A (en) | Generating capacity demand prediction method based on LSTM deep learning | |
| CN117010946A (en) | Thermal power plant production and operation cost accounting system and application method thereof | |
| Zhang et al. | Ensemble learning-based approach for residential building heating energy prediction and optimization | |
| CN111027816B (en) | Photovoltaic power generation efficiency calculation method based on data envelope analysis | |
| CN110874611B (en) | Wind power cluster power ultra-short term prediction error grade grading compensation method | |
| Sičaja et al. | Evaluation of load forecast model performance in Croatian DSO |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |