CN109165774A - A kind of short-term photovoltaic power prediction technique - Google Patents
A kind of short-term photovoltaic power prediction technique Download PDFInfo
- Publication number
- CN109165774A CN109165774A CN201810889441.XA CN201810889441A CN109165774A CN 109165774 A CN109165774 A CN 109165774A CN 201810889441 A CN201810889441 A CN 201810889441A CN 109165774 A CN109165774 A CN 109165774A
- Authority
- CN
- China
- Prior art keywords
- day
- prediction
- vector
- state
- data
- 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 42
- 238000013528 artificial neural network Methods 0.000 claims abstract description 34
- 238000004458 analytical method Methods 0.000 claims abstract description 24
- 238000012549 training Methods 0.000 claims abstract description 23
- 238000010248 power generation Methods 0.000 claims abstract description 17
- 238000013459 approach Methods 0.000 claims abstract description 9
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000012546 transfer Methods 0.000 claims description 7
- 102000008297 Nuclear Matrix-Associated Proteins Human genes 0.000 claims description 3
- 108010035916 Nuclear Matrix-Associated Proteins Proteins 0.000 claims description 3
- 210000000299 nuclear matrix Anatomy 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims description 3
- 238000011084 recovery Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 abstract description 4
- 230000002411 adverse Effects 0.000 abstract description 3
- 239000000203 mixture Substances 0.000 abstract description 3
- 239000003016 pheromone Substances 0.000 description 13
- 241000257303 Hymenoptera Species 0.000 description 9
- 239000000284 extract Substances 0.000 description 5
- 238000007418 data mining Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000005611 electricity Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000013277 forecasting method Methods 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 230000003068 static effect Effects 0.000 description 3
- 210000004027 cell Anatomy 0.000 description 2
- 238000010219 correlation analysis Methods 0.000 description 2
- 230000008034 disappearance Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 235000013305 food Nutrition 0.000 description 2
- 230000006386 memory function Effects 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 241001269238 Data Species 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 210000004218 nerve net Anatomy 0.000 description 1
- 210000004940 nucleus Anatomy 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
-
- 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)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Biophysics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- General Business, Economics & Management (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Marketing (AREA)
- Computational Linguistics (AREA)
- Tourism & Hospitality (AREA)
- Molecular Biology (AREA)
- Data Mining & Analysis (AREA)
- Water Supply & Treatment (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Public Health (AREA)
- Entrepreneurship & Innovation (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Photovoltaic Devices (AREA)
Abstract
The invention discloses a kind of short-term photovoltaic power prediction technique, content includes: collection historical data, including power data and relevant weather data, obtains training sample time series, line number of going forward side by side Data preprocess；The history power generation day composition preliminary sample of identical season and day weather pattern is picked out from sample database according to the season of prediction day and weather pattern forecast information；It is determined using gray relative analysis method and predicts day meteorological immediate several similar days；Photovoltaic data characteristics is extracted based on core principle component analysis method；Using the weight and threshold value of ant group algorithm training neural network, photovoltaic power is predicted using trained neural network；The predicted value tentatively obtained is modified using Markov approach, obtains final prediction result.The present invention improves precision of prediction compared with prior art, alleviates the grid-connected caused adverse effect of photovoltaic generating system, ensures that the planning and stable operation of entire electric system.
Description
Technical field
The invention belongs to generations of electricity by new energy and intelligent power grid technology field, and in particular to a kind of to use correlation rule, core master
The hybrid forecasting method of constituent analysis and Elman neural network predicts short-term photovoltaic power.
Background technique
With the continuous social and economic development, the mankind are also increasing for the dependence of the energy and demand, and energy problem gets over
More to become the bottleneck for restricting human social development.Solar energy because its is inexhaustible, nexhaustible and does not pollute characteristic at
For most ideal and with the fastest developing speed one of green energy resource.As the important way that solar energy development utilizes, photovoltaic power generation
Have many advantages, such as that no fuel consumption, non-pollutant discharge, application form are flexible, there is boundless application prospect.However, light
Volt power generation has fluctuation and intermittence, this gesture vulnerable to protean intensity of solar radiation and such environmental effects, output
Must seriously affect its it is grid-connected after, the stability and reliability of electric system.Therefore, it is badly in need of finding a kind of accurately and reliably photovoltaic
Generated power forecasting method, to mitigate the grid-connected caused adverse effect of photovoltaic generating system, it is ensured that the rule of entire electric system
Draw and stable operation.
The core of Accurate Prediction is the processing to historical sample data, however these data are often very huge,
It is an important link of prediction that reasonable effective data, which how therefrom to be chosen, as training sample.Grey correlation analysis is as number
According to one of digging technology method, complexity is low, has higher treatment effeciency to large data sets.Use it as processing history
The method of data can extract reasonable effective sample data.
In view of the influence factor of photovoltaic output power is more, there is certain correlations between each variable, with this data
As training sample, it is difficult to reach higher precision of prediction.In order to remove between input variable there is redundancy, principal component
Analysis (principal component analysis, PCA) method is widely adopted.However, it only relates to the original in data
Beginning spatially carries out linearization process.Compared with this method, core principle component analysis (kernel PCA, KPCA) method can be effective
Ground carries out the Nonlinear feature extraction of data on higher-dimension nucleus lesion, thus improve extract nonlinear transformations it is comprehensive,
Validity and the speed of service.
In the selection of forecasting tool, traditional static feedforward neural network method can only realize that static non linear maps at present
Function is easily trapped into local minimum；Support vector machines is sensitive to the selection of parameter, and training speed is slow.In order to overcome
Disadvantages mentioned above, J.L.Elman propose the Elman nerve net with Dynamic Recurrent performance for speech processing problems in nineteen ninety
Network, the network can approach arbitrary function within the limited time, and there has also been very big promotions for training speed.
Photovoltaic power generation fluctuation is larger, is the random process of a non-stationary, the neural network structure of finite sample training
Often imperfect stability.Since state probability transfer matrix has the ability of tracking variable random fluctuation, it is used for pair
The prediction result of neural network is corrected, and the precision of prediction of prediction model can be improved.
Summary of the invention
It is an object of that present invention to provide a kind of short-term photovoltaic power prediction technique, this method be based on data mining, core it is main at
Analysis and Elman neural network carry out hybrid predicting, and prediction accuracy is higher.
To achieve the above object, the invention adopts the following technical scheme:
A kind of short-term photovoltaic power prediction technique, this method are based on data mining, core principle component analysis and Elman nerve
The hybrid forecasting method of network, to improve precision of prediction.Wherein grey correlation analysis is a kind of data digging method, it is by sentencing
The correlation degree of each factor in disconnected system, extracts several history generated output days with prediction day highlights correlations, and composition has height
The data sequence of similar features is spent as training sample；In order to remove existing redundancy between input variable, core principle component point
Analysis method be used to carry out feature extraction to training sample, replace original a large amount of inputs with less input；Elman neural network conduct
A kind of accurate predictive tools with dynamic memory function be used to carry out tentative prediction to photovoltaic power；Finally for prediction
Model predict that error is larger at the peak value of power swing and precision of prediction there are fluctuations, using Markov approach to preliminary
Obtained predicted value is modified, and obtains final prediction result.
The content of prediction technique of the present invention includes the following steps:
Step 1: historical data, including power data and relevant weather data are collected, training sample time series is obtained, and
Carry out data prediction；
Step 2: identical season and day are picked out from sample database according to the season of prediction day and weather pattern forecast information
The history power generation day of weather pattern forms preliminary sample；
Step 3: being determined using gray relative analysis method and predict day meteorological immediate several similar days；
Step 4: photovoltaic data characteristics is extracted based on core principle component analysis method；
Step 5: using the weight and threshold value of ant group algorithm training neural network, using trained neural network to photovoltaic
Power is predicted；
Step 6: the predicted value tentatively obtained being modified using Markov approach, obtains final prediction result.
Further, described to be determined and prediction day meteorological immediate several phases using gray relative analysis method in step 3
It is exactly the determining and prediction day immediate several similar days of meteorology on the basis of step 2 simple classification like day, it is specific interior
Appearance includes the following steps:
Step 3.1: choose daily Meteorological Characteristics vector:
xi=[thi,t1,tli,t2,tai,rh,ri] (1)
Wherein, thi, tli, taiRespectively i-th day daily maximum temperature, the lowest temperature and temperature on average, unit are DEG C；t1
And t2At the time of respectively daily maximum temperature and the lowest temperature occur；Rh is relative humidity, is indicated with % (percentage)；riFor day
Rainfall, unit mm；
Step 3.2: the Meteorological Characteristics vector of note prediction day is X=[x (1), x (2) ..., x (n)]T, then i-th of history day
Meteorologic factor feature vector be denoted as Xi=[xi(1),xi(2),…,xi(n)]T, wherein n is characterized the number of component of a vector；
To i-th of history day, k-th of Meteorological Characteristics component of a vector is normalized:
Wherein, xiIt (k) is k-th of Meteorological Characteristics component of a vector of i-th of history day；xi,max(k) and xi,min(k) it is respectively
The maximum value and minimum value of k-th of Meteorological Characteristics component of a vector；
Step 3.3: calculating the incidence coefficient of prediction day and i-th of history day, k-th of Meteorological Characteristics component of a vector are as follows:
Wherein, x'(k) and x'i(k) prediction day and k-th on the i-thth Meteorological Characteristics component of a vector after respectively normalizing；
ρ is resolution ratio, usually takes ρ=0.5；
Step 3.4: calculating i-th and predict the total correlation degree of day are as follows:
Since closest to history day, calculate the similarity with prediction day one by one, and in order of dates arrange, choose phase
Like degree FiSimilar day of a days of >=0.85 as prediction day.
Further, described that photovoltaic data characteristics is extracted based on core principle component analysis method in step 4, it is exactly according to core master
Componential analysis extracts photovoltaic data characteristics, the specific steps of which are as follows:
Step 4.1: for the generated power forecasting of daily N number of sampled point, it is assumed that total M sample, then it is available initial
Input variable matrix XM×N；
Step 4.2: calculating the eigenvalue λ of nuclear matrix K after centralizationiWith corresponding Orthogonal Units feature vector ei；
Step 4.3: after calculating characteristic value, being arranged according to numerical values recited, such as λ1≥λ2≥…≥λM；
Step 4.4: calculating the contribution rate η of characteristic valueiAnd contribution rate of accumulative total ηΣ(i), principal component number and master are determined
Ingredient component Yi。
Further, described to utilize ant group algorithm training neural network weight and threshold value, weight and threshold value in step 5
Calculating steps are as follows:
Step 5.1: setting primary condition；
Step 5.2: starting all ants, every ant k successively exists since the 1st set according to path finding algorithm
An element is selected in each set；
Step 5.3: step 5.2 is repeated, until all ants in ant colony all arrive at food source；
Step 5.4: when all ants have selected an element in each set, and ant nest is returned to according to original route,
If the process undergoes n chronomere, then the pheromones of selected element adjust as the following formula:
τj(Ipi) (t+n)=ρ τj(Ipi)(t)+Δτj(Ipi) (5)
In formula, parameter ρ (0 < ρ < 1) indicates the residual degree of pheromones, then 1- ρ indicates the disappearance degree of pheromones；Kth ant is indicated in this circulation, in set IpiJ-th of element pj(Ipi) on the pheromones that leave, can use
Following formula calculates:
In formula, Q is constant, for adjusting regulating the speed for pheromones, ekIt is to make one group of element of kth ant selection
For neural network weight and threshold value when, maximum output error after training sample set:
In formula, S is number of samples, OnAnd OexIt is the reality output and desired output of neural network respectively, therefore ekIt is smaller, phase
Answer the increase of pheromones more；
Step 5.5: when all ants all converge to a paths or the number of iterations N >=Nmax, then iteration terminates, output
Optimal solution；Otherwise 5.2 are gone to step.
Further, described that the predicted value tentatively obtained is modified using Markov approach in step 6, it obtains
Final prediction result, the specific steps of which are as follows:
Step 6.1: according to the predicted value of neural network, calculating Relative Error δ；
Step 6.2: calculating and work as state EiState E is transferred to after k timesjWhen probability
Step 6.3: using lower threshold value in difference locating for the relative error δ of test sample predicted value as state demarcation value
State demarcation standard is established in domain；According to residual error state computation state transition probability matrix；
Step 6.4: the error information predicted according to photovoltaic power generation prediction model determines initial state vector X (0), and count
The state for calculating kth step shifts result；
Step 6.5: according to obtained Prediction of Markov relative error, calculating revised photovoltaic power value P.
The present invention provides existing redundancies between a kind of consideration photovoltaic generation power influence factor, and prediction mould
Type tentative prediction result there are the accurately and reliably prediction technique of error, alleviate photovoltaic generating system it is grid-connected caused by it is unfavorable
It influences, ensures that the planning and stable operation of entire electric system.
Due to the adoption of the above technical scheme, a kind of short-term photovoltaic power prediction technique provided by the invention, with the prior art
Compared to have it is such the utility model has the advantages that
1, the present invention preferably can excavate effective information from a large amount of historical datas by the use of data mining algorithm,
And gray relative analysis method uses so that whole process very time-saving and efficiency.
2, present invention introduces core principle component analysis methods, and to carry out dimensionality reduction to the input number of training sample preferred, after making dimensionality reduction
Data dependence reduces, and improves precision of prediction.
3, for the present invention using having the Elman neural network of dynamic memory function as forecasting tool, this method overcomes biography
Static feedforward neural network training speed of uniting is slow and the disadvantage of generalization ability difference, further improves precision of prediction.
4, most of prediction is all directly using the output of model as final prediction result, and present invention introduces Ma Erke
Husband's chain compensates and corrects the tentative prediction error of model, obtains final prediction result, improves precision of prediction.
Detailed description of the invention
Fig. 1 is the flow diagram of the prediction technique；
Fig. 2 is the structure chart of forecasting tool Elman neural network of the present invention；
Fig. 3 is the Relative Error comparison diagram of fine day of the present invention；
Fig. 4 is fine day 24-h photovoltaic prediction result of the present invention；
Fig. 5 is the Relative Error comparison diagram of rainy day of the invention；
Fig. 6 is rainy day 24-h photovoltaic prediction result of the invention.
Specific embodiment
Present invention is further described in detail with specific embodiment with reference to the accompanying drawing:
A kind of short-term photovoltaic power prediction technique of the invention, since the randomness and fluctuation of photovoltaic power generation are larger, directly
Prediction effect variation certainly will be will lead to by connecing prediction, first be replaced the exceptional value in former historical data in the method for the present invention, and draw
There is the data mining technology of unique advantage used in processing big data quantity aspect, find and predict the multiple of day same weather category
Historical power day form the data sequence with height similar features；Secondly, using core principle component analysis method to training sample
It is preferred to input number progress dimensionality reduction, extracts chief composition series；Then, the Elman neural network pair optimized using ant group algorithm
Data are predicted；Finally, predicting that error is larger and pre- at the peak value of power swing for Elman neural network prediction model
Surveying precision, there are fluctuations, are modified using Markov approach to the predicted value tentatively obtained, obtain final prediction knot
Fruit.The flow chart such as Fig. 1 institute predicted using the Elman neural network short-term photovoltaic power of correlation rule and core principle component analysis
Show, the specific steps of which are as follows:
Step 1, historical data, including power data and relevant weather data are collected, training sample time series is obtained, and
Carry out data prediction.
Firstly, usually containing many unreasonable data beyond normal range (NR) in historical data, these data can be right
Prediction causes adverse effect, therefore substitutes these data using mean value method:
Wherein, pi,mean(t)、pi-1(t) and pi-2(t) be respectively i-th, (i-1)-th and the i-th -2 days same t moments light
Lie prostrate generated output.
Step 2, the major influence factors for determining photovoltaic system electricity generation power, it is pre- according to the season of prediction day and weather pattern
Breath of notifying picks out the history power generation day of identical season and day weather pattern from sample database, forms preliminary sample:
The photovoltaic power generation system output power calculation formula of unit area are as follows:
ps=η SI [1-0.005 (t0+25)] (10)
Wherein, η is photovoltaic cell transfer efficiency, is indicated with % (percentage)；S is photovoltaic array area, unit m2；I
For solar irradiation intensity, unit kW/m2；t0For environment temperature, unit is DEG C.
According to formula (10) it is found that influence photovoltaic system electricity generation power principal element be weather pattern, solar irradiation intensity,
Environment temperature, the transfer efficiency of photovoltaic cell and array area.For same photovoltaic generating system, the influence factors such as η and S are
Included in history power generation data, do not consider further that.The Weather information that meteorological department provides includes weather pattern, temperature, wet substantially
Degree and wind-force.
Identical season and day weather class are picked out from sample database according to the season of prediction day and weather pattern forecast information
The history power generation day of type forms preliminary sample.
Step 3, it determines using gray relative analysis method and predicts day meteorological immediate several similar days, specific steps are such as
Under:
(1) it is as follows to choose daily Meteorological Characteristics vector:
xi=[thi,t1,tli,t2,tai,rh,ri] (11)
Wherein, thi, tli, taiRespectively i-th day daily maximum temperature, the lowest temperature and temperature on average, unit are DEG C；t1,
t2At the time of respectively daily maximum temperature and the lowest temperature occur；Rh is relative humidity, is indicated with % (percentage)；riFor day drop
Rainfall, unit mm；
(2) the Meteorological Characteristics vector of note prediction day is X=[x (1), x (2) ..., x (n)]T, then the gas of i-th of history day
As factor feature vector is denoted as Xi=[xi(1),xi(2),…,xi(n)]T, wherein n is characterized the number of component of a vector；
To i-th of history day, k-th of Meteorological Characteristics component of a vector is normalized:
Wherein, xiIt (k) is k-th of Meteorological Characteristics component of a vector of i-th of history day；xi,max(k) and xi,min(k) it is respectively
The maximum value and minimum value of k-th of Meteorological Characteristics component of a vector；
(3) incidence coefficient of prediction day and i-th of history day, k-th of Meteorological Characteristics component of a vector is calculated are as follows:
Wherein, x'(k) and x'i(k) prediction day and k-th on the i-thth Meteorological Characteristics component of a vector after respectively normalizing；
ρ is resolution ratio, usually takes ρ=0.5；
(4) calculate i-th and predict the total correlation degree of day are as follows:
Since closest to history day, calculate the similarity with prediction day one by one, and in order of dates arrange, choose phase
Like degree FiSimilar day of a days of >=0.85 as prediction day.
Step 4, photovoltaic data characteristics is extracted according to core principle component analysis method, the specific steps are as follows:
(1) for the generated power forecasting of daily N number of sampled point, it is assumed that total M sample, then available initial input becomes
Moment matrix XM×N, it may be assumed that
(2) to input variable matrix XM×NPrincipal component search be equal to be calculate centralization after nuclear matrix K characteristic value
λiWith corresponding Orthogonal Units feature vector ei；
(3) it after calculating characteristic value, is arranged according to numerical values recited, such as λ1≥λ2≥…≥λM；
(4) principal component number and principal component component Y are determined according to the recovery rate of settingi；Foundation is to calculate characteristic value
Contribution rate ηiAnd contribution rate of accumulative total ηΣ(i):
Step 5, using the weight and threshold value of ant group algorithm training neural network, using trained neural network to photovoltaic
Power is predicted.
Wherein the ant group algorithm training neural network weight and threshold value, steps are as follows for the calculating of weight and threshold value:
(1) primary condition is set: enabling time t=0 and the number of iterations N=0, setting maximum number of iterations is Nmax, ant
Number is l, enables set IpiThe pheromones τ that each element in (1≤i≤m) is carved at the beginningj(Ipi) (0)=0, and Δ τj
(IpiWhole ants are placed in ant nest by)=0；
(2) start all ants, every ant k is since the 1st set, according to following path finding algorithms successively every
An element is selected in a set；
Path finding algorithm: for set Ipi, any one ant k (k=1,2 ..., l) be random according to the probability P of calculating
Ground selects its j-th of element:
(3) step (2) are repeated, until all ants in ant colony all arrive at food source；
(4) element has been selected in each set when all ants, and has returned to ant nest according to original route, if the mistake
Cheng Jingli n chronomere, then the pheromones of selected element adjust as the following formula:
τj(Ipi) (t+n)=ρ τj(Ipi)(t)+Δτj(Ipi) (20)
In formula, parameter ρ (0 < ρ < 1) indicates the residual degree of pheromones, then 1- ρ indicates the disappearance degree of pheromones；Kth ant is indicated in this circulation, in set IpiJ-th of element pj(Ipi) on the pheromones that leave, can
It is calculated with following formula:
In formula, Q is constant, for adjusting regulating the speed for pheromones, ekIt is to make one group of element of kth ant selection
For neural network weight and threshold value when, maximum output error after training sample set:
In formula, S is number of samples, OnAnd OexIt is the reality output and desired output of neural network respectively, therefore ekIt is smaller, phase
Answer the increase of pheromones more；
(5) when all ants all converge to a paths or the number of iterations N >=Nmax, then iteration terminates, and exports optimal
Solution.Otherwise (2) are gone to step.
Step 6, the predicted value tentatively obtained is modified using Markov approach, obtains final prediction result,
The specific steps of which are as follows:
(1) according to the predicted value of neural network, Relative Error δ is calculated:
In formula, PexFor neural network prediction generated energy, PnFor actual power generation；
(2) as state EiState E is transferred to after k timesjWhen, probability are as follows:
In formula,It is sample state from EiTo EjTransfer number, AiThe total degree occurred for state；
(3) it using lower threshold value in difference locating for the relative error δ of test sample predicted value as state demarcation codomain, establishes
State demarcation standard；According to residual error state computation state transition probability matrix P(k):
In formula, n is the state number after classification；
(4) error information predicted according to photovoltaic power generation prediction model, determines initial state vector X (0), and is turned by state
Move the state transfer result that formula (27) calculate kth step:
X (k)=X (0) P(k) (27)
(5) according to obtained Prediction of Markov relative error, revised photovoltaic power value P is calculated:
In formula, δh、δlFor the upper lower threshold value in locating error state section.
Below by example, the present invention will be described, and the data of photovoltaic power generation come from U.S.'s new energy laboratory here.
By taking fine day as an example, photovoltaic power prediction is carried out to it.It determines using gray relative analysis method first and predicts day similarity highest
5 days, related coefficient is respectively as follows: 0.93,0.92,0.90,0.87,0.85 from high to low.Next, choosing influences photovoltaic hair
14 major influence factors of electrical power: to prediction the highest 5 days synchronizations of day similarity performance number, with predict day it is similar
It is the performance number at 2 moment before and after spending highest 1 day moment, the maximum temperature for predicting day, minimum temperature, mean temperature, opposite
Humidity, rainfall predict temperature, the solar irradiation intensity at moment day.Then, this obtained 14 major influence factors are adopted
Dimension-reduction treatment is carried out with KPCA method, and kernel function uses Radial basis kernel function, the characteristic value being calculated, variance contribution ratio and tired
Meter variance contribution ratio is listed in following table 1:
Table 1
Taking the contribution rate of accumulative total of variance upper limit is 95%, as it can be seen from table 1 the contribution rate of accumulative total of variance of preceding 3 principal components
Reach 95.4707%, illustrates that this 3 principal components can replace 14 original influence factors.
The principal component after feature will be extracted to be predicted as training set using Elman neural network, and to tentatively obtaining
Prediction result carry out Markov residual GM, obtained prediction result is compared with other methods.It is this shown in Fig. 2
The structure chart of forecasting tool Elman neural network used in inventing.Method 1 is the method for the present invention, and method 2 is without Ma Erke
The prediction technique of husband's residual GM, method 3 are using the sample extracted without core principle component as Elman neural network input quantity
Prediction technique.Three kinds of prediction techniques are as shown in Figure 3 and Figure 5 in the Relative Error comparison diagram of fine day and rainy day.Fine day and
Two kinds of rainy day 24-h representative photovoltaic power prediction result curves are as shown in Figure 4 and Figure 6.It can be seen that the present invention predicts
Method can effectively improve precision of prediction.
Claims (4)
1. a kind of short-term photovoltaic power prediction technique, it is characterised in that: the particular content of the prediction technique includes the following steps:
Step 1: historical data, including power data and relevant weather data are collected, training sample time series is obtained, goes forward side by side
Line number Data preprocess；
Step 2: identical season and day weather are picked out from sample database according to the season of prediction day and weather pattern forecast information
The history power generation day of type forms preliminary sample；
Step 3: it is determined using gray relative analysis method and predicts day meteorological immediate several similar days；
Step 4: photovoltaic data characteristics is extracted based on core principle component analysis method；
Step 5: using the weight and threshold value of ant group algorithm training neural network, using trained neural network to photovoltaic function
Rate is predicted；
Step 6: the predicted value tentatively obtained is modified using Markov approach, obtains final prediction result.
2. a kind of short-term photovoltaic power prediction technique according to claim 1, it is characterised in that: described in step 3
It is determined using gray relative analysis method and predicts day meteorological immediate several similar days, be exactly the base in step 2 simple classification
On plinth, determines and include the following steps: with prediction day meteorological immediate several similar days, particular content
Step 1: choose daily Meteorological Characteristics vector:
xi=[thi,t1,tli,t2,tai,rh,ri] (1)
Wherein, thi, tli, taiRespectively i-th day daily maximum temperature, the lowest temperature and temperature on average, unit are DEG C；t1And t2Point
It Wei not daily maximum temperature and the lowest temperature at the time of occur；Rh is relative humidity, is expressed as a percentage；riIt is single for daily rainfall
Position is mm；
Step 2: the Meteorological Characteristics vector of note prediction day is X=[x (1), x (2) ..., x (n)]T, then the meteorology of i-th of history day
Factor feature vector is denoted as Xi=[xi(1),xi(2),…,xi(n)]T, wherein n is characterized the number of component of a vector；
To i-th of history day, k-th of Meteorological Characteristics component of a vector is normalized:
Wherein, xiIt (k) is k-th of Meteorological Characteristics component of a vector of i-th of history day；xi,max(k) and xi,minIt (k) is respectively kth
The maximum value and minimum value of a Meteorological Characteristics component of a vector；
Step 3: calculating the incidence coefficient of prediction day and i-th of history day, k-th of Meteorological Characteristics component of a vector are as follows:
Wherein, x'(k) and x 'i(k) prediction day and k-th on the i-thth Meteorological Characteristics component of a vector after respectively normalizing；ρ is point
It distinguishes coefficient, usually takes ρ=0.5；
Step 4: calculating i-th and predict the total correlation degree of day are as follows:
Since closest to history day, calculate the similarity with prediction day one by one, and in order of dates arrange, choose similarity
FiSimilar day of a days of >=0.85 as prediction day.
3. a kind of short-term photovoltaic power prediction technique according to claim 1, it is characterised in that: described in step 4
Photovoltaic data characteristics is extracted based on core principle component analysis method, is exactly that photovoltaic data characteristics is extracted according to core principle component analysis method,
Specific step is as follows:
Step 1: for the generated power forecasting of daily N number of sampled point, it is assumed that total M sample, then available initial input becomes
Moment matrix XM×N, it may be assumed that
Step 2: to input variable matrix XM×NPrincipal component search be equal to be calculate centralization after nuclear matrix K characteristic value
λiWith corresponding Orthogonal Units feature vector ei；
Step 3: after calculating characteristic value, being arranged according to numerical values recited, such as λ1≥λ2≥…≥λM；
Step 4: principal component number and principal component component Y are determined according to the recovery rate of settingi；According to the tribute for being calculating characteristic value
Offer rate ηiAnd contribution rate of accumulative total ηΣ(i):
4. a kind of short-term photovoltaic power prediction technique according to claim 1, it is characterised in that: described in step 6
The predicted value tentatively obtained is modified using Markov approach, obtains final prediction result, the specific steps of which are as follows:
Step 1: according to the predicted value of neural network, calculate Relative Error δ:
In formula, PexFor neural network prediction generated energy, PnFor actual power generation；
Step 2: as state EiState E is transferred to after k timesjWhen, probability are as follows:
In formula,It is sample state from EiTo EjTransfer number, AiThe total degree occurred for state；
Step 3: using lower threshold value in difference locating for the relative error δ of test sample predicted value as state demarcation codomain, establishing
State demarcation standard；According to residual error state computation state transition probability matrix P(k):
In formula, n is the state number after classification；
Step 4: the error information predicted according to photovoltaic power generation prediction model determines initial state vector X (0), and is turned by state
Move the state transfer result that formula (12) calculate kth step:
X (k)=X (0) P(k) (12)
Step 5: according to obtained Prediction of Markov relative error, calculate revised photovoltaic power value P:
In formula, δh、δlFor the upper lower threshold value in locating error state section.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810889441.XA CN109165774A (en) | 2018-08-07 | 2018-08-07 | A kind of short-term photovoltaic power prediction technique |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810889441.XA CN109165774A (en) | 2018-08-07 | 2018-08-07 | A kind of short-term photovoltaic power prediction technique |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109165774A true CN109165774A (en) | 2019-01-08 |
Family
ID=64895200
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810889441.XA Pending CN109165774A (en) | 2018-08-07 | 2018-08-07 | A kind of short-term photovoltaic power prediction technique |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109165774A (en) |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110751326A (en) * | 2019-10-17 | 2020-02-04 | 江苏远致能源科技有限公司 | Photovoltaic day-ahead power prediction method and device and storage medium |
| CN110796292A (en) * | 2019-10-14 | 2020-02-14 | 国网辽宁省电力有限公司盘锦供电公司 | Photovoltaic power short-term prediction method considering haze influence |
| CN110889564A (en) * | 2019-12-16 | 2020-03-17 | 吉林大学 | Short-term power load prediction method based on MKPCA-RBFNN |
| CN111626473A (en) * | 2020-04-17 | 2020-09-04 | 国网上海市电力公司 | Two-stage photovoltaic power prediction method considering error correction |
| CN111695601A (en) * | 2020-05-15 | 2020-09-22 | 特变电工西安电气科技有限公司 | Photovoltaic power prediction method, device, equipment and readable storage medium |
| CN111985678A (en) * | 2020-07-06 | 2020-11-24 | 上海交通大学 | Photovoltaic power short-term prediction method |
| CN112183849A (en) * | 2020-09-25 | 2021-01-05 | 贵州乌江水电开发有限责任公司 | Short-term electric power quantity prediction method based on artificial neural network |
| JP6833303B1 (en) * | 2019-12-10 | 2021-02-24 | 東芝三菱電機産業システム株式会社 | Power generation forecaster |
| CN112418502A (en) * | 2020-11-13 | 2021-02-26 | 国网冀北电力有限公司计量中心 | Photovoltaic power generation prediction method based on combined model |
| CN112488421A (en) * | 2020-12-15 | 2021-03-12 | 国网雄安金融科技集团有限公司 | Tracking and predicting method and device for electric charge account receivable |
| CN112529285A (en) * | 2020-12-08 | 2021-03-19 | 湖南工业大学 | Photovoltaic power generation power prediction method based on similar daily theory and PCA-PSO-BP |
| CN112927097A (en) * | 2021-01-29 | 2021-06-08 | 国网辽宁省电力有限公司阜新供电公司 | Photovoltaic power generation short-term prediction method based on GRA-ABC-Elman model |
| CN112949913A (en) * | 2021-02-08 | 2021-06-11 | 国网河北省电力有限公司衡水供电分公司 | Photovoltaic output prediction method and terminal equipment |
| CN113446988A (en) * | 2021-06-08 | 2021-09-28 | 武汉理工大学 | Airport runway pavement state monitoring system and method based on cloud edge fusion framework |
| CN113723670A (en) * | 2021-08-12 | 2021-11-30 | 浙江大学 | Photovoltaic power generation power short-term prediction method with variable time window |
| CN113961614A (en) * | 2021-09-06 | 2022-01-21 | 中国农业大学 | Photovoltaic power prediction method based on similar day selection and classified time-sharing error correction |
| CN114048896A (en) * | 2021-10-27 | 2022-02-15 | 国核自仪系统工程有限公司 | Method, system, equipment and medium for predicting photovoltaic power generation data |
| CN115800385A (en) * | 2022-08-15 | 2023-03-14 | 国网安徽省电力有限公司经济技术研究院 | Electric energy quality regulation and control method based on adjustable and controllable capacity of photovoltaic inverter and charging pile |
| CN116702937A (en) * | 2022-12-14 | 2023-09-05 | 国网湖北省电力有限公司荆门供电公司 | Photovoltaic output day-ahead prediction method based on K-means mean value clustering and BP neural network optimization |
| CN116911418A (en) * | 2022-12-14 | 2023-10-20 | 国网湖北省电力有限公司荆门供电公司 | Photovoltaic power generation power ultra-short-term prediction method based on wavelet transformation and BP neural network optimization |
| CN116911467A (en) * | 2023-09-12 | 2023-10-20 | 浙江华云电力工程设计咨询有限公司 | Renewable energy output prediction method, device and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105160423A (en) * | 2015-09-14 | 2015-12-16 | 河海大学常州校区 | Photovoltaic power generation prediction method based on Markov residual error correction |
| CN106447098A (en) * | 2016-09-22 | 2017-02-22 | 许昌许继软件技术有限公司 | Photovoltaic ultra-short period power predicting method and device |
| CN106570594A (en) * | 2016-11-11 | 2017-04-19 | 国网江西省电力公司电力科学研究院 | Similar day photovoltaic power generation short period prediction method based on TMBP |
| CN106600041A (en) * | 2016-12-02 | 2017-04-26 | 天津大学 | Restricted Boltzmann machine-based photovoltaic power generation short-term power probability prediction method |
-
2018
- 2018-08-07 CN CN201810889441.XA patent/CN109165774A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105160423A (en) * | 2015-09-14 | 2015-12-16 | 河海大学常州校区 | Photovoltaic power generation prediction method based on Markov residual error correction |
| CN106447098A (en) * | 2016-09-22 | 2017-02-22 | 许昌许继软件技术有限公司 | Photovoltaic ultra-short period power predicting method and device |
| CN106570594A (en) * | 2016-11-11 | 2017-04-19 | 国网江西省电力公司电力科学研究院 | Similar day photovoltaic power generation short period prediction method based on TMBP |
| CN106600041A (en) * | 2016-12-02 | 2017-04-26 | 天津大学 | Restricted Boltzmann machine-based photovoltaic power generation short-term power probability prediction method |
Non-Patent Citations (1)
| Title |
|---|
| CHUNXIA DOU,ETC.: "Elman neural network based short-term photovoltaic power forecasting using association rules and kernel principal component analysis", 《JOURNAL OF RENEWABLE AND SUSTAINABLE ENERGY》 * |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110796292A (en) * | 2019-10-14 | 2020-02-14 | 国网辽宁省电力有限公司盘锦供电公司 | Photovoltaic power short-term prediction method considering haze influence |
| CN110751326A (en) * | 2019-10-17 | 2020-02-04 | 江苏远致能源科技有限公司 | Photovoltaic day-ahead power prediction method and device and storage medium |
| JP6833303B1 (en) * | 2019-12-10 | 2021-02-24 | 東芝三菱電機産業システム株式会社 | Power generation forecaster |
| WO2021117127A1 (en) * | 2019-12-10 | 2021-06-17 | 東芝三菱電機産業システム株式会社 | Power generation amount estimation device |
| CN110889564A (en) * | 2019-12-16 | 2020-03-17 | 吉林大学 | Short-term power load prediction method based on MKPCA-RBFNN |
| CN111626473A (en) * | 2020-04-17 | 2020-09-04 | 国网上海市电力公司 | Two-stage photovoltaic power prediction method considering error correction |
| CN111695601B (en) * | 2020-05-15 | 2023-06-20 | 特变电工西安电气科技有限公司 | Photovoltaic power prediction method, device, equipment and readable storage medium |
| CN111695601A (en) * | 2020-05-15 | 2020-09-22 | 特变电工西安电气科技有限公司 | Photovoltaic power prediction method, device, equipment and readable storage medium |
| CN111985678A (en) * | 2020-07-06 | 2020-11-24 | 上海交通大学 | Photovoltaic power short-term prediction method |
| CN112183849A (en) * | 2020-09-25 | 2021-01-05 | 贵州乌江水电开发有限责任公司 | Short-term electric power quantity prediction method based on artificial neural network |
| CN112418502A (en) * | 2020-11-13 | 2021-02-26 | 国网冀北电力有限公司计量中心 | Photovoltaic power generation prediction method based on combined model |
| CN112418502B (en) * | 2020-11-13 | 2024-04-30 | 国网冀北电力有限公司计量中心 | Photovoltaic power generation prediction method based on combined model |
| CN112529285A (en) * | 2020-12-08 | 2021-03-19 | 湖南工业大学 | Photovoltaic power generation power prediction method based on similar daily theory and PCA-PSO-BP |
| CN112488421B (en) * | 2020-12-15 | 2023-04-28 | 国网雄安金融科技集团有限公司 | Tracking and predicting method and device for accounts receivable of electric charge |
| CN112488421A (en) * | 2020-12-15 | 2021-03-12 | 国网雄安金融科技集团有限公司 | Tracking and predicting method and device for electric charge account receivable |
| CN112927097A (en) * | 2021-01-29 | 2021-06-08 | 国网辽宁省电力有限公司阜新供电公司 | Photovoltaic power generation short-term prediction method based on GRA-ABC-Elman model |
| CN112949913A (en) * | 2021-02-08 | 2021-06-11 | 国网河北省电力有限公司衡水供电分公司 | Photovoltaic output prediction method and terminal equipment |
| CN113446988A (en) * | 2021-06-08 | 2021-09-28 | 武汉理工大学 | Airport runway pavement state monitoring system and method based on cloud edge fusion framework |
| CN113723670A (en) * | 2021-08-12 | 2021-11-30 | 浙江大学 | Photovoltaic power generation power short-term prediction method with variable time window |
| CN113723670B (en) * | 2021-08-12 | 2023-12-08 | 浙江大学 | Photovoltaic power generation power short-term prediction method with variable time window |
| CN113961614A (en) * | 2021-09-06 | 2022-01-21 | 中国农业大学 | Photovoltaic power prediction method based on similar day selection and classified time-sharing error correction |
| CN113961614B (en) * | 2021-09-06 | 2024-10-15 | 中国农业大学 | Photovoltaic power prediction method based on similar day selection and classification time-sharing error correction |
| CN114048896A (en) * | 2021-10-27 | 2022-02-15 | 国核自仪系统工程有限公司 | Method, system, equipment and medium for predicting photovoltaic power generation data |
| CN114048896B (en) * | 2021-10-27 | 2023-02-03 | 国核自仪系统工程有限公司 | Method, system, equipment and medium for predicting photovoltaic power generation data |
| CN115800385B (en) * | 2022-08-15 | 2024-04-19 | 国网安徽省电力有限公司经济技术研究院 | Power quality regulation method based on capacity regulation of photovoltaic inverter and charging pile |
| CN115800385A (en) * | 2022-08-15 | 2023-03-14 | 国网安徽省电力有限公司经济技术研究院 | Electric energy quality regulation and control method based on adjustable and controllable capacity of photovoltaic inverter and charging pile |
| CN116911418A (en) * | 2022-12-14 | 2023-10-20 | 国网湖北省电力有限公司荆门供电公司 | Photovoltaic power generation power ultra-short-term prediction method based on wavelet transformation and BP neural network optimization |
| CN116702937A (en) * | 2022-12-14 | 2023-09-05 | 国网湖北省电力有限公司荆门供电公司 | Photovoltaic output day-ahead prediction method based on K-means mean value clustering and BP neural network optimization |
| CN116911467A (en) * | 2023-09-12 | 2023-10-20 | 浙江华云电力工程设计咨询有限公司 | Renewable energy output prediction method, device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109165774A (en) | A kind of short-term photovoltaic power prediction technique | |
| Liu et al. | Forecasting power output of photovoltaic system using a BP network method | |
| CN109873610B (en) | Photovoltaic array fault diagnosis method based on IV characteristic and depth residual error network | |
| CN108549929B (en) | A kind of photovoltaic power prediction technique based on deep layer convolutional neural networks | |
| CN109002915B (en) | Photovoltaic power station short-term power prediction method based on Kmeans-GRA-Elman model | |
| CN106447098B (en) | Photovoltaic ultra-short-term power prediction method and device | |
| CN110705789A (en) | Photovoltaic power station short-term power prediction method | |
| CN108205717A (en) | A kind of photovoltaic generation power Multiple Time Scales Forecasting Methodology | |
| CN107194495B (en) | Photovoltaic power longitudinal prediction method based on historical data mining | |
| CN111260126B (en) | Short-term photovoltaic power generation prediction method considering correlation degree of weather and meteorological factors | |
| CN107609697B (en) | A kind of Wind power forecasting method | |
| CN108280545A (en) | A kind of photovoltaic power prediction technique based on K mean cluster neural network | |
| CN114792156A (en) | Photovoltaic output power prediction method and system based on curve characteristic index clustering | |
| CN110866633B (en) | Micro-grid ultra-short-term load prediction method based on SVR support vector regression | |
| CN110766134A (en) | Photovoltaic power station short-term power prediction method based on cyclic neural network | |
| CN112418346B (en) | Numerical weather forecast total radiation system error classification calculation method | |
| CN112070311A (en) | Day-ahead light power prediction method based on similar day clustering and meteorological factor weighting | |
| CN114462718A (en) | CNN-GRU wind power prediction method based on time sliding window | |
| CN116345555A (en) | CNN-ISCA-LSTM model-based short-term photovoltaic power generation power prediction method | |
| CN109242180A (en) | Long-medium term power load forecasting method and system | |
| CN104102951A (en) | Short-term wind power prediction method based on EMD historical data preprocessing | |
| CN116629416A (en) | Photovoltaic power station power prediction method and device | |
| CN115759389A (en) | Day-ahead photovoltaic power prediction method based on weather type similar day combination strategy | |
| CN116128150A (en) | Photovoltaic intelligent output prediction method based on two-stage optimization | |
| Choulli et al. | Hybrid optimization based on the analytical approach and the particle swarm optimization algorithm (Ana-PSO) for the extraction of single and double diode models parameters |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190108 |