CN116911467A - Renewable energy output prediction method, device and storage medium - Google Patents
Renewable energy output prediction method, device and storage medium Download PDFInfo
- Publication number
- CN116911467A CN116911467A CN202311166833.0A CN202311166833A CN116911467A CN 116911467 A CN116911467 A CN 116911467A CN 202311166833 A CN202311166833 A CN 202311166833A CN 116911467 A CN116911467 A CN 116911467A
- Authority
- CN
- China
- Prior art keywords
- period
- influencing
- factors
- factor set
- 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 93
- 230000004913 activation Effects 0.000 claims abstract description 18
- 239000011159 matrix material Substances 0.000 claims description 37
- 239000013598 vector Substances 0.000 claims description 30
- 238000011176 pooling Methods 0.000 claims description 16
- 238000000605 extraction Methods 0.000 claims description 11
- 230000002159 abnormal effect Effects 0.000 claims description 9
- 238000010586 diagram Methods 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 5
- 230000003213 activating effect Effects 0.000 claims description 3
- 238000004590 computer program Methods 0.000 claims description 3
- 239000000758 substrate Substances 0.000 claims 2
- 238000004364 calculation method Methods 0.000 abstract description 12
- 230000007774 longterm Effects 0.000 abstract description 4
- 238000012958 reprocessing Methods 0.000 abstract description 4
- 238000013527 convolutional neural network Methods 0.000 description 36
- 230000006870 function Effects 0.000 description 17
- 238000012545 processing Methods 0.000 description 7
- 238000010248 power generation Methods 0.000 description 6
- 238000012549 training Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 238000005286 illumination Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000002123 temporal effect Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 238000012886 linear function Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 239000002803 fossil fuel Substances 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000000053 physical method Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012706 support-vector machine 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
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Operations Research (AREA)
- Primary Health Care (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application relates to a renewable energy output prediction method, a device and a storage medium, which are applied to the technical field of renewable energy output prediction and comprise the following steps: according to the application, a main influence factor set of reproducible output is constructed, similar time intervals are screened according to the main influence factor set, ESN and CNN architectures with residual connection are integrated, wherein time information is extracted by combining ESN, spatial information is extracted by combining CNN network, and then the output of a mixed model is taken as a basis to carry out final prediction, so that the performance is improved, the ESN and CNN are connected through residual, the problem of vanishing gradient in characteristic representation and long-term time dependence thereof can be solved by reprocessing the previous activation function, and compared with a single method or other mixed methods, the method provides better prediction precision and lower time complexity, remarkably reduces error rate and has lower calculation complexity.
Description
Technical Field
The application relates to the technical field of renewable energy output prediction, in particular to a renewable energy output prediction method, a renewable energy output prediction device and a storage medium.
Background
Renewable energy power generation is considered one of the most promising solutions compared to fossil fuel power generation because it has key features of green, clean and natural complement over a wide geographic area. However, the use of renewable energy also brings non-dispatchable uncertainty, which reduces the stability and reliability of the energy sector, especially for large scale grid-connected renewable energy sources, where renewable energy sources have randomness, intermittence and strong volatility, which increases the cost of power generation, while also requiring a large number of electronic devices, thereby reducing the moment of inertia and stability of the power system.
Accurate renewable energy output prediction (REPP) is an effective method for reducing uncertainty, which is very important for operation, management and planning of energy departments and systems, and compared with traditional physical and statistical methods, artificial intelligence-based prediction models produce a pleasing result because they have potential in extracting representative features from historical data, however, these models including support vector machines, back propagation algorithms, maximum entropy and lifting methods employ shallow architectures developed in the 80 s, the main problem of these models is the need to artificially set feature engineering, partial generalization capability and sample complexity, and due to the huge RE data, shallow models can suffer from parameter non-convergence and calculation instability, and Deep Learning (DL) has unique unsupervised learning, big data training and powerful generalization capability compared with shallow models.
The current deep learning prediction is mainly performed by a single method, for example, a model based on CNN is very robust in terms of extracting spatial information, but the historical generating capacity data is time series data and contains spatial and temporal information, and the single model is poor in prediction accuracy.
Disclosure of Invention
In view of the above, the present application aims to provide a method, a device and a storage medium for predicting renewable energy output, so as to solve the problems that in the prior art, prediction is performed by a single CNN model, extraction of spatial information is more robust, but historical generating capacity data is time series data, and simultaneously comprises spatial and time information, and the single model is poor in prediction accuracy.
According to a first aspect of an embodiment of the present application, there is provided a method for predicting output of renewable energy, the method including:
acquiring historical period sample data of wind power and photovoltaic output power, and acquiring influence factors influencing the wind power and the photovoltaic unit output power according to the historical period sample data;
inputting the acquired influence factors into a preset ISM model to obtain a multi-order data matrix, solving the multi-order data matrix, and acquiring a main influence factor set influencing wind power and output power of a photovoltaic unit;
selecting a sample closest to each parameter data of a period to be predicted from the historical period sample data as a similar period sample, and acquiring influence factors in a main influence factor set of the similar period sample;
inputting influencing factors in the main influencing factor set of the similar period sample into an ESN model, wherein the ESN model outputs time characteristics of the influencing factors in the main influencing factor set of the similar period sample;
and inputting the time feature into a convolution layer of a preset CNN model, carrying out convolution on the time feature and a convolution kernel of the CNN model by the convolution layer of the CNN model to output a feature vector, sequentially repeating convolution pooling operation on the feature vector, extracting spatial features until a final pooling layer outputs a feature map, and predicting wind power and photovoltaic output power of a period to be predicted through the feature map to obtain an output power predicted value.
Preferably, the method further comprises:
after the historical period sample data of wind power and photovoltaic output power are obtained, removing abnormal values from the historical period sample data through a three-sigma rule of thumb;
filling missing values in the historical period sample data after the abnormal values are removed by a NAN interpolation method;
and converting the historical period sample data subjected to missing value filling into standard historical period sample data in a specific range by adopting a minimum-maximum data normalization method.
Preferably, the method comprises the steps of,
inputting the obtained influence factors into a preset ISM model to obtain a multi-order data matrix, solving the multi-order data matrix, and obtaining a main influence factor set influencing the output power of the wind power unit and the output power of the photovoltaic unit comprises the following steps:
the obtained influencing factors are used as input data and are input into an ISM model to formOrder data matrix->；
Calculating the data matrixMiddle->The correlation coefficient between the term influence factors is obtained>Order coefficient matrix；
By presetting a threshold valueFor coefficient matrix->Processing to obtain/>Order adjacency matrix->；
Adjacency matrix pair through preset operation rulesTreating to obtain->Order reachable matrix->；
Obtaining a reachability matrixThe method comprises the steps of obtaining a PQS set according to a reachable set P, a preceding set Q and an intersection S of the reachable set P, the preceding set Q and the preceding set Q corresponding to each element;
in the PQS set, the index is screened from the highest layer of influencing factors, and before selectionThe item index is used as a result index group; screening indexes from the bottommost layer, before selecting +.>And taking the item index as a driving index group, and taking the collection of the result index group and the driving index group as a main influencing factor set for influencing the output power of the wind power unit and the photovoltaic unit.
Preferably, the method comprises the steps of,
inputting the influence factors in the main influence factor set of the similar period sample into an ESN model, and outputting the time characteristics of the influence factors in the main influence factor set of the similar period sample by the ESN model comprises the following steps:
inputting influencing factors in the main influencing factor set of the similar period samples to an input layer of an ESN model;
sequentially repeating the activation operation on the influence factors in the main influence factor set of the samples in the similar period in the storage layer through the activation function of the storage layer of the ESN model, and transmitting the influence factors to the output layer of the ESN model;
and repeatedly activating the influence factors in the main influence factor set of the similar period sample through the activation function of the output layer of the ESN model to finish the extraction of the time characteristics of the influence factors in the main influence factor set of the similar period sample.
Preferably, the method further comprises:
acquiring the actual value of wind power and photovoltaic output power in a period to be predicted;
and obtaining an error index through the output power predicted value and the actual value, and judging the accuracy of the predicted result through the error index.
Preferably, the method comprises the steps of,
the selecting, from the historical period sample data, a sample closest to each parameter data of the period to be predicted as a similar period sample includes:
acquiring a judging feature vector of each period of the history through judging factors of each period in the sample data of the history period;
acquiring a judging feature vector of the period to be predicted through judging factors of the period to be predicted;
calculating the association degree of each historical period and the period to be predicted respectively through judging the feature vector;
and selecting the historical time period with the highest association degree as a similar time period sample.
According to a second aspect of an embodiment of the present application, there is provided a device for predicting output of renewable energy, the device comprising:
the influence factor acquisition module is used for: the method comprises the steps of acquiring historical period sample data of wind power and photovoltaic output power, and acquiring influence factors influencing the wind power and the photovoltaic unit output power according to the historical period sample data;
the main influence factor set acquisition module: the method comprises the steps of inputting the acquired influence factors into a preset ISM model to obtain a multi-order data matrix, solving the multi-order data matrix, and acquiring a main influence factor set influencing wind power and output power of a photovoltaic unit;
a similar period acquisition module: the method comprises the steps of selecting a sample closest to each parameter data of a period to be predicted from historical period sample data as a similar period sample, and acquiring influence factors in a main influence factor set of the similar period sample;
and the time feature extraction module is used for: inputting influencing factors in the main influencing factor set of the similar period sample into an ESN model, wherein the ESN model outputs time characteristics of the influencing factors in the main influencing factor set of the similar period sample;
and a prediction module: the method comprises the steps that the time features are input into a convolution layer of a preset CNN model, the convolution layer of the CNN model carries out convolution on the time features and a convolution kernel of the CNN model to output feature vectors, the feature vectors are sequentially and repeatedly subjected to convolution pooling operation to extract spatial features until a last pooling layer outputs a feature diagram, and wind power and photovoltaic output power of a period to be predicted are predicted through the feature diagram to obtain an output power predicted value.
According to a third aspect of embodiments of the present application, there is provided a storage medium storing a computer program which, when executed by a master, implements the steps of the method of predicting renewable energy output.
The technical scheme provided by the embodiment of the application can comprise the following beneficial effects:
according to the application, a main influence factor set of reproducible output is constructed, similar time periods are screened according to the main influence factor set, ESN and CNN architectures with residual connection are integrated, wherein time information is extracted by combining ESN, spatial information is extracted by combining CNN network, then ISM-ESN-CNN output is used as a basis for final prediction, so that performance is improved, ESN and CNN are connected through residual connection, the problem of vanishing gradient in characteristic representation and long-term time dependence thereof can be solved by reprocessing the previous activation function, and compared with a single method or other mixed methods, the method provides better prediction precision and lower time complexity, remarkably reduces error rate and has lower calculation complexity.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application as claimed.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and together with the description, serve to explain the principles of the application.
FIG. 1 is a flow chart illustrating a method of predicting renewable energy output, according to an exemplary embodiment;
FIG. 2 is a system diagram illustrating a device for predicting renewable energy output, according to another exemplary embodiment;
in the accompanying drawings: the system comprises a 1-influence factor acquisition module, a 2-main influence factor set acquisition module, a 3-similar period acquisition module, a 4-time feature extraction module and a 5-prediction module.
Detailed Description
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the application. Rather, they are merely examples of apparatus and methods consistent with aspects of the application as detailed in the accompanying claims.
Example 1
FIG. 1 is a flow chart illustrating a method for predicting renewable energy output, according to an exemplary embodiment, as shown in FIG. 1, the method comprising:
s1, acquiring historical period sample data of wind power and photovoltaic output power, and acquiring influence factors influencing the wind power and the photovoltaic unit output power according to the historical period sample data;
s2, inputting the acquired influence factors into a preset ISM model to obtain a multi-order data matrix, solving the multi-order data matrix, and acquiring a main influence factor set influencing the output power of the wind power unit and the output power of the photovoltaic unit;
s3, selecting a sample closest to each parameter data of the period to be predicted from the historical period sample data as a similar period sample, and acquiring influence factors in a main influence factor set of the similar period sample;
s4, inputting the influence factors in the main influence factor set of the similar period sample into an ESN model, and outputting the time characteristics of the influence factors in the main influence factor set of the similar period sample by the ESN model;
s5, inputting the time feature into a convolution layer of a preset CNN model, carrying out convolution on the time feature and a convolution kernel of the CNN model by the convolution layer of the CNN model to output a feature vector, sequentially repeating convolution pooling operation on the feature vector, extracting spatial features until a feature map is output by a last pooling layer, and predicting wind power and photovoltaic output power to be predicted in a period of time through the feature map to obtain an output power predicted value;
it can be understood that, obtaining historical period sample data of wind power and photovoltaic output power, including according to historical wind power data, unit states, weather conditions and the like, a wind farm is generally powered by a plurality of wind turbines of the same type in a cluster according to a certain rule, and the output power expression of a single wind turbine is as follows:
in the method, in the process of the application,the output power of the wind turbine generator is; />Is air density; />Is the wind energy utilization coefficient; />Is the blade size; />Is the actual wind speed; />Rated output power of the wind turbine generator; />The cut-in wind speed of the wind turbine generator is the cut-in wind speed of the wind turbine generator; />The rated wind speed of the wind turbine generator is set; />The wind speed is cut out for the wind turbine generator;
the photovoltaic module unit can directly convert solar energy into electric energy, the output of the photovoltaic module unit is stable on sunny days, and the output power can be expressed as follows:
in the method, in the process of the application,is the actual output force of the photovoltaic module, +.>Is photoelectric conversion efficiency, < >>、/>And->Respectively atmospheric temperature, normal operating temperature and reference temperature, +.>Representing the intensity of the optical radiation at time t +.>Is the reference efficiency of the photovoltaic panel, < >>Refers to the area of a single photovoltaic panel, +.>Refers to the number of photovoltaic panels;
according to the analysis of the output power of the wind power generation unit and the photovoltaic unit, factors such as wind speed, wind direction, temperature, humidity, air pressure, tip speed ratio, pitch angle and the like influence the wind power output power, and factors such as efficiency, temperature, radiation intensity, photovoltaic area and quantity influence the photovoltaic output power; taking the influence factors of the wind and light output power considered in the steps as input data, inputting the input data into an ISM model to obtain a multi-order data matrix, solving the multi-order data matrix, and obtaining a main influence factor set influencing the output power of wind power and photovoltaic units; based on the main influence factor set established in the steps, selecting a period which is most similar to a period to be predicted according to historical wind power data, unit states, weather conditions and the like, and forming a training sample and a test sample of output power, wherein the similar period is a sample which is selected from historical period samples of wind power photovoltaic output power and is most similar to each parameter data of the period to be predicted; establishing an ESN model, performing time learning through the ESN model, wherein the ESN is a special and novel form of RNN, is easier to realize and is more intelligent in calculation than other RNN variants, the ESN provides complete framework and supervision learning strategy for the RNN, the ESN consists of an input layer, a storage layer and an output layer, wherein the input unit I is distributed to the input layer, the storage layer is provided with an internal unit R, the output layer is provided with an output unit O, and the screened wind speed, wind direction, illumination intensity, air temperature, weather and other data of similar time periods are transmittedIs added to a reserve layer, the time characteristics of the influencing factors in the main influencing factor set of the similar period sample are output through an ESN model, and the data obtained by the ESN are transmittedGiven that CNNs, which are special neural networks specifically designed for capturing local (spatial) information from input data, by explicitly encoding them in an architecture, CNNs can better process image (2D) data by expanding 1D or 3D data, CNN architectures have two large features of weight sharing and local connectivity, which makes them different from conventional neural networks, the main architecture of CNN models generally includes a convolution layer, where the convolution operation is the main idea of CNNs, for obtaining local spatial features from data, and convolving the temporal features output by the ESN model with its own convolution kernel in the convolution layer to output feature vectorsXWhich may include a plurality of eigenvector convolutions, if +.>Is a eigenvector and n is the number of samples, the mathematical operation of the first convolution layer may be performed as follows:
wherein,,is the output of the first convolution layer,Xis a feature vector, < >>Is the deviation, W represents the weight index, m represents the index filter value, < >>Is an activation function;
first, thelThe output of the layer convolution layer can be calculated by:
wherein a ReLU activation function is used, as follows:
and transmitting the data obtained by the ESN to the CNN, and sequentially repeating convolution pooling operation until the last pooling layer outputs a characteristic diagram, so that the wind power can be predicted. According to the application, a main influence factor set of reproducible output is constructed, similar time periods are screened according to the main influence factor set, ESN and CNN architectures with residual connection are integrated, wherein time information is extracted by combining ESN, spatial information is extracted by combining CNN network, then ISM-ESN-CNN output is used as a basis for final prediction, so that performance is improved, ESN and CNN are connected through residual connection, the problem of vanishing gradient in characteristic representation and long-term time dependence thereof can be solved by reprocessing the previous activation function, and compared with a single method or other mixed methods, the method provides better prediction precision and lower time complexity, remarkably reduces error rate and has lower calculation complexity.
Preferably, the method further comprises:
after the historical period sample data of wind power and photovoltaic output power are obtained, removing abnormal values from the historical period sample data through a three-sigma rule of thumb;
filling missing values in the historical period sample data after the abnormal values are removed by a NAN interpolation method;
converting the historical period sample data subjected to missing value filling into standard historical period sample data in a specific range by adopting a minimum-maximum data normalization method;
it will be appreciated that in general, the hybrid model performs better than a single model in terms of renewable energy output prediction, and therefore, the present application develops a hybrid model based on a combination of ESN and CNN models for extracting temporal and spatial features from historical data of power production, first preprocessing an input sequence of historical power generation data to remove abnormal data such as abnormal values, missing values, and the like, and removing the abnormal values using a three sigma rule of thumb, the mathematical expression of which is as follows:
where f is a function of,is a data set->Representing data values of renewable power generation; />Anddata set +.>Average and standard deviation of (a);
to fill in the missing values, we use the NAN interpolation method, whose mathematical expression is as follows:
wherein nan represents a missing set of values;
after removing outliers and filling in missing values, we then apply a min-max data normalization procedure to convert the refined data into a specific range, because renewable energy production data is diverse in nature, neural networks are more sensitive to such data ranges, and the mathematical representation of the min-max normalization procedure is as follows:
and then the refined data is forwarded to the model for learning prediction.
Preferably, the method comprises the steps of,
inputting the obtained influence factors into a preset ISM model to obtain a multi-order data matrix, solving the multi-order data matrix, and obtaining a main influence factor set influencing the output power of the wind power unit and the output power of the photovoltaic unit comprises the following steps:
the obtained influencing factors are used as input data and are input into an ISM model to formOrder data matrix->；
Calculating the data matrixMiddle->The correlation coefficient between the term influence factors is obtained>Order coefficient matrix；
By presetting a threshold valueFor coefficient matrix->Treating to obtain->Order adjacency matrix->；
Adjacency matrix pair through preset operation rulesTreating to obtain->Order reachable matrix->；
Obtaining reachabilityMatrix arrayThe method comprises the steps of obtaining a PQS set according to a reachable set P, a preceding set Q and an intersection S of the reachable set P, the preceding set Q and the preceding set Q corresponding to each element;
in the PQS set, the index is screened from the highest layer of influencing factors, and before selectionThe item index is used as a result index group; screening indexes from the bottommost layer, before selecting +.>The item index is used as a driving index group, and the collection of the result index group and the driving index group is used as a main influencing factor set for influencing the output power of the wind power unit and the photovoltaic unit;
it can be understood that the influence factors of the wind-solar output power considered in the above steps are taken as input data and input into an ISM model to formOrder data matrix->To properly reduce the influence index screening range, namely:
in the method, in the process of the application,a value representing the nth index of the mth group of data,/->For index number, ->The number of the data groups;
for data matrixProcessing and calculating data matrix->Middle->The correlation coefficient between the term influence factors is obtained>Order coefficient matrix->To->For example, the first and second columns of the table are as follows:
obtainingOrder coefficient matrix->：
In the method, in the process of the application,is the association coefficient;
according to a preset threshold valueFor coefficient matrix->Treating to obtain->Order adjacency matrix->The processing mode is as follows: if->Let->Otherwise let->Wherein->The value can be defined, and is generally 0.85, and the obtained valueOrder adjacency matrix->The method comprises the following steps:
pair of adjacency matricesTreating to obtain->Order reachable matrix->Wherein the adjacency matrix->And reachability matrix->The following operation rules are satisfied:
for reachable matrixProcessing to find +.>The reachable set P, the antecedent set Q and the intersection S of the reachable set P and the antecedent set Q corresponding to each element in the list. The reachable set P is solved by finding out the column corresponding to the element 1 in each row; solving the preceding set Q is to find out the row corresponding to 1 element in each column, obtain the reachable set P corresponding to each element, the preceding set Q and the intersection S of the reachable set P and the preceding set Q by processing, so as to form a PQS set, screen and optimize main influencing factors on the basis of the obtained PQS set, screen indexes from the highest layer of influencing factors, and select the preceding one>The item index is used as a result index group; screening indexes from the bottommost layer, before selecting +.>And the item indexes are used as driving index groups, output result index groups and driving index groups, and main influencing factors are screened out to form a main influencing factor set of wind power output power.
Preferably, the method comprises the steps of,
inputting the influence factors in the main influence factor set of the similar period sample into an ESN model, and outputting the time characteristics of the influence factors in the main influence factor set of the similar period sample by the ESN model comprises the following steps:
inputting influencing factors in the main influencing factor set of the similar period samples to an input layer of an ESN model;
sequentially repeating the activation operation on the influence factors in the main influence factor set of the samples in the similar period in the storage layer through the activation function of the storage layer of the ESN model, and transmitting the influence factors to the output layer of the ESN model;
repeatedly activating the influence factors in the main influence factor set of the similar period sample through an activation function of an output layer of the ESN model to finish the extraction of the time characteristics of the influence factors in the main influence factor set of the similar period sample;
it will be appreciated that the ESN consists of an input layer to which the input unit I is assigned, a storage layer with an internal unit R, and an output layer with an output unit O, inputu) Inside is [ ]x) And output unity) The calculation formulas of (a) are respectively as follows:
a typical update formula for the internal unit and the output unit is as follows:
the data of the influence factors in the similar period of wind speed, wind direction, illumination intensity, air temperature, weather and other main influence factors are imported into the input layer, the operation is sequentially repeated in the storage layer through the activation function, then the data of the storage layer is transferred to the output layer, the operation is repeatedly performed through the activation function of the output layer, and the ESN realizes the extraction of the time characteristics of the characteristic factors; in the actual prediction process, the method mainly comprises the following steps:
(1) initializing and setting related parameters;
(2) wind speed, wind direction and illumination intensity of the screened similar periodData such as degree, temperature, weather, etcIs added to the reserve layer;
(3) the calculation of the state variables and the output variables of the reserve layer is sequentially completed according to the typical updating formulas of the internal unit and the output unit, and the matrix form is formed, and the reading function g of the output layer can be linearly read by using a linear function or can be non-linearly read by using a non-linear function;
(4) according to the output weightCalculation formula output +.>Said->The calculation formula of (2) is as follows:
in the above-mentioned method, the step of,fandgthe activation functions of the storage layer unit and the output layer unit are respectively represented, and the total weight index，/>，And->Respectively representing input, reservoir, return and output weights, M representing the state matrix of the reservoir,/->In practice, it is usually omitted that the weight matrix from the storage layer is +.>Is updated in the learning process, and the weight index is randomly selected and kept unchanged, and the weight value is returned +.>Set to 0, neglecting feedback of the output layer to the dynamic reservoir, thus training the ESN, namely the output weight +.>Training is performed.
Preferably, the method further comprises:
acquiring the actual value of wind power and photovoltaic output power in a period to be predicted;
obtaining an error index through the output power predicted value and the actual value, and judging the accuracy of the predicted result through the error index;
it will be appreciated that the performance of the model may be assessed on the basis of MSE, MBE, MAE and RMSE metrics, which are widely used in previous model assessments. MSE is the difference between the actual and estimated values of the model, while MBE is the average deviation of the actual and predicted values of the model. Negative and positive values of MBE represent model underestimation and overestimation, respectively, MAE is the average of the absolute values between the test set and the model output values, and RMSE is the square root mean difference, the mathematical expression of these indices is as follows:
wherein,,is the actual value; />Is the value of the model prediction.
Preferably, the method comprises the steps of,
the selecting, from the historical period sample data, a sample closest to each parameter data of the period to be predicted as a similar period sample includes:
acquiring a judging feature vector of each period of the history through judging factors of each period in the sample data of the history period;
acquiring a judging feature vector of the period to be predicted through judging factors of the period to be predicted;
calculating the association degree of each historical period and the period to be predicted respectively through judging the feature vector;
selecting a history period with highest association degree as a similar period sample;
it can be understood that the similar period refers to a sample closest to each parameter data of the period to be measured, which is selected from samples of the historical period of the wind power photovoltaic output power, and a feature vector is constructed by a judgment factor of the similar period, and wind power is taken as an example here, namely:
in the method, in the process of the application,representing a decision feature vector; />Respectively representing the maximum value, the minimum value and the average value of the wind speed in the wind power output power history period; />Respectively representing sine values and cosine values of wind directions in a history period; />Average temperature representing historical time period; />Average humidity representing the historical period; />Representing the average air pressure over the history period.
Time period to test and thThe decision feature vectors of the individual history periods are respectively:
in the method, in the process of the application,a determination feature vector representing a period to be measured; />Indicate->Determination features of individual history periods
Vector;represents the ∈th of the period to be measured>Determining factors; />Indicate->The first historical period
A step of judgment;
by the extremely bad methodFor a pair ofAnd->Carrying out normalization dimension processing to reduce calculation difficulty and error;
calculating the association coefficient and the association degree of each historical period and the period to be measured;
determining a similar time period range, sorting the historical time periods according to the relevance calculation result in descending order,
namely:selecting->Is a similar period.
Example two
FIG. 2 is a system diagram illustrating a renewable energy power prediction device according to another exemplary embodiment, including:
influence factor acquisition module 1: the method comprises the steps of acquiring historical period sample data of wind power and photovoltaic output power, and acquiring influence factors influencing the wind power and the photovoltaic unit output power according to the historical period sample data;
the main influencing factor set acquisition module 2: the method comprises the steps of inputting the acquired influence factors into a preset ISM model to obtain a multi-order data matrix, solving the multi-order data matrix, and acquiring a main influence factor set influencing wind power and output power of a photovoltaic unit;
the similar period acquisition module 3: the method comprises the steps of selecting a sample closest to each parameter data of a period to be predicted from historical period sample data as a similar period sample, and acquiring influence factors in a main influence factor set of the similar period sample;
time feature extraction module 4: inputting influencing factors in the main influencing factor set of the similar period sample into an ESN model, wherein the ESN model outputs time characteristics of the influencing factors in the main influencing factor set of the similar period sample;
prediction module 5: the method comprises the steps that the time characteristics are input into a convolution layer of a preset CNN model, the convolution layer of the CNN model carries out convolution on the time characteristics and a convolution kernel of the CNN model to output characteristic vectors, the characteristic vectors are sequentially and repeatedly subjected to convolution pooling operation to extract spatial characteristics until a last pooling layer outputs a characteristic diagram, wind power and photovoltaic output power of a period to be predicted are predicted through the characteristic diagram, and an output power predicted value is obtained;
it can be understood that the influence factor obtaining module 1 is used for obtaining historical period sample data of wind power and photovoltaic output power, and obtaining influence factors influencing the wind power and the photovoltaic unit output power according to the historical period sample data; the method comprises the steps that a main influence factor set acquisition module 2 is used for inputting acquired influence factors into a preset ISM model to obtain a multi-order data matrix, solving the multi-order data matrix and acquiring a main influence factor set influencing output power of a wind power unit and a photovoltaic unit; the similar period acquisition module 3 is used for selecting a sample closest to each parameter data of the period to be predicted from the historical period sample data as a similar period sample, and acquiring influence factors in a main influence factor set of the similar period sample; the time feature extraction module 4 is used for inputting the influence factors in the main influence factor set of the similar period sample into an ESN model, and the ESN model outputs the time features of the influence factors in the main influence factor set of the similar period sample; the prediction module 5 is used for inputting the time feature into a convolution layer of a preset CNN model, the convolution layer of the CNN model carries out convolution on the time feature and a convolution kernel of the CNN model to output a feature vector, the feature vector is sequentially and repeatedly subjected to convolution pooling operation, spatial feature extraction is carried out until a final pooling layer outputs a feature map, wind power and photovoltaic output power of a period to be predicted are predicted through the feature map, and an output power predicted value is obtained; according to the application, a main influence factor set of reproducible output is constructed, similar time periods are screened according to the main influence factor set, ESN and CNN architectures with residual connection are integrated, wherein time information is extracted by combining ESN, spatial information is extracted by combining CNN network, then ISM-ESN-CNN output is used as a basis for final prediction, so that performance is improved, ESN and CNN are connected through residual connection, the problem of vanishing gradient in characteristic representation and long-term time dependence thereof can be solved by reprocessing the previous activation function, and compared with a single method or other mixed methods, the method provides better prediction precision and lower time complexity, remarkably reduces error rate and has lower calculation complexity.
Embodiment III:
the present embodiment provides a storage medium storing a computer program which, when executed by a master controller, implements each step in the above method;
it is to be understood that the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.
It is to be understood that the same or similar parts in the above embodiments may be referred to each other, and that in some embodiments, the same or similar parts in other embodiments may be referred to.
It should be noted that in the description of the present application, the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. Furthermore, in the description of the present application, unless otherwise indicated, the meaning of "plurality" means at least two.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and further implementations are included within the scope of the preferred embodiment of the present application in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present application.
It is to be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, the various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, may be implemented using any one or combination of the following techniques, as is well known in the art: discrete logic circuits having logic gates for implementing logic functions on data signals, application specific integrated circuits having suitable combinational logic gates, programmable Gate Arrays (PGAs), field Programmable Gate Arrays (FPGAs), and the like.
Those of ordinary skill in the art will appreciate that all or a portion of the steps carried out in the method of the above-described embodiments may be implemented by a program to instruct related hardware, where the program may be stored in a computer readable storage medium, and where the program, when executed, includes one or a combination of the steps of the method embodiments.
In addition, each functional unit in the embodiments of the present application may be integrated in one processing module, or each unit may exist alone physically, or two or more units may be integrated in one module. The integrated modules may be implemented in hardware or in software functional modules. The integrated modules may also be stored in a computer readable storage medium if implemented in the form of software functional modules and sold or used as a stand-alone product.
The above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, or the like.
In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
While embodiments of the present application have been shown and described above, it will be understood that the above embodiments are illustrative and not to be construed as limiting the application, and that variations, modifications, alternatives and variations may be made to the above embodiments by one of ordinary skill in the art within the scope of the application.
Claims (8)
1. A method for predicting renewable energy output, the method comprising:
acquiring historical period sample data of wind power and photovoltaic output power, and acquiring influence factors influencing the wind power and the photovoltaic unit output power according to the historical period sample data;
inputting the acquired influence factors into a preset ISM model to obtain a multi-order data matrix, solving the multi-order data matrix, and acquiring a main influence factor set influencing wind power and output power of a photovoltaic unit;
selecting a sample closest to each parameter data of a period to be predicted from the historical period sample data as a similar period sample, and acquiring influence factors in a main influence factor set of the similar period sample;
inputting influencing factors in the main influencing factor set of the similar period sample into an ESN model, wherein the ESN model outputs time characteristics of the influencing factors in the main influencing factor set of the similar period sample;
and inputting the time feature into a convolution layer of a preset CNN model, carrying out convolution on the time feature and a convolution kernel of the CNN model by the convolution layer of the CNN model to output a feature vector, sequentially repeating convolution pooling operation on the feature vector, extracting spatial features until a final pooling layer outputs a feature map, and predicting wind power and photovoltaic output power of a period to be predicted through the feature map to obtain an output power predicted value.
2. The method as recited in claim 1, further comprising:
after the historical period sample data of wind power and photovoltaic output power are obtained, removing abnormal values from the historical period sample data through a three-sigma rule of thumb;
filling missing values in the historical period sample data after the abnormal values are removed by a NAN interpolation method;
and converting the historical period sample data subjected to missing value filling into standard historical period sample data in a specific range by adopting a minimum-maximum data normalization method.
3. The method of claim 1, wherein the step of determining the position of the substrate comprises,
inputting the obtained influence factors into a preset ISM model to obtain a multi-order data matrix, solving the multi-order data matrix, and obtaining a main influence factor set influencing the output power of the wind power unit and the output power of the photovoltaic unit comprises the following steps:
the obtained influencing factors are used as input data and are input into an ISM model to formOrder data matrix->；
Calculating the data matrixMiddle->The correlation coefficient between the term influence factors is obtained>Order coefficient matrix->；
By presetting a threshold valueFor coefficient matrix->Treating to obtain->Order adjacency matrix->；
Adjacency matrix pair through preset operation rulesTreating to obtain->Order reachable matrix->；
Obtaining a reachability matrixThe method comprises the steps of obtaining a PQS set according to a reachable set P, a preceding set Q and an intersection S of the reachable set P, the preceding set Q and the preceding set Q corresponding to each element;
in the PQS set, the index is screened from the highest layer of influencing factors, and before selectionThe item index is used as a result index group; screening indexes from the bottommost layer, before selecting +.>And taking the item index as a driving index group, and taking the collection of the result index group and the driving index group as a main influencing factor set for influencing the output power of the wind power unit and the photovoltaic unit.
4. The method of claim 3, wherein the step of,
inputting the influence factors in the main influence factor set of the similar period sample into an ESN model, and outputting the time characteristics of the influence factors in the main influence factor set of the similar period sample by the ESN model comprises the following steps:
inputting influencing factors in the main influencing factor set of the similar period samples to an input layer of an ESN model;
sequentially repeating the activation operation on the influence factors in the main influence factor set of the samples in the similar period in the storage layer through the activation function of the storage layer of the ESN model, and transmitting the influence factors to the output layer of the ESN model;
and repeatedly activating the influence factors in the main influence factor set of the similar period sample through the activation function of the output layer of the ESN model to finish the extraction of the time characteristics of the influence factors in the main influence factor set of the similar period sample.
5. The method as recited in claim 1, further comprising:
acquiring the actual value of wind power and photovoltaic output power in a period to be predicted;
and obtaining an error index through the output power predicted value and the actual value, and judging the accuracy of the predicted result through the error index.
6. The method of claim 1, wherein the step of determining the position of the substrate comprises,
the selecting, from the historical period sample data, a sample closest to each parameter data of the period to be predicted as a similar period sample includes:
acquiring a judging feature vector of each period of the history through judging factors of each period in the sample data of the history period;
acquiring a judging feature vector of the period to be predicted through judging factors of the period to be predicted;
calculating the association degree of each historical period and the period to be predicted respectively through judging the feature vector;
and selecting the historical time period with the highest association degree as a similar time period sample.
7. A device for predicting renewable energy output, the device comprising:
the influence factor acquisition module is used for: the method comprises the steps of acquiring historical period sample data of wind power and photovoltaic output power, and acquiring influence factors influencing the wind power and the photovoltaic unit output power according to the historical period sample data;
the main influence factor set acquisition module: the method comprises the steps of inputting the acquired influence factors into a preset ISM model to obtain a multi-order data matrix, solving the multi-order data matrix, and acquiring a main influence factor set influencing wind power and output power of a photovoltaic unit;
a similar period acquisition module: the method comprises the steps of selecting a sample closest to each parameter data of a period to be predicted from historical period sample data as a similar period sample, and acquiring influence factors in a main influence factor set of the similar period sample;
and the time feature extraction module is used for: inputting influencing factors in the main influencing factor set of the similar period sample into an ESN model, wherein the ESN model outputs time characteristics of the influencing factors in the main influencing factor set of the similar period sample;
and a prediction module: the method comprises the steps that the time features are input into a convolution layer of a preset CNN model, the convolution layer of the CNN model carries out convolution on the time features and a convolution kernel of the CNN model to output feature vectors, the feature vectors are sequentially and repeatedly subjected to convolution pooling operation to extract spatial features until a last pooling layer outputs a feature diagram, and wind power and photovoltaic output power of a period to be predicted are predicted through the feature diagram to obtain an output power predicted value.
8. A storage medium storing a computer program which, when executed by a master, performs the steps of a method of predicting a renewable energy output according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311166833.0A CN116911467A (en) | 2023-09-12 | 2023-09-12 | Renewable energy output prediction method, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311166833.0A CN116911467A (en) | 2023-09-12 | 2023-09-12 | Renewable energy output prediction method, device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116911467A true CN116911467A (en) | 2023-10-20 |
Family
ID=88356856
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311166833.0A Pending CN116911467A (en) | 2023-09-12 | 2023-09-12 | Renewable energy output prediction method, device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116911467A (en) |
Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105373857A (en) * | 2015-11-27 | 2016-03-02 | 许昌许继软件技术有限公司 | Photovoltaic power station irradiance prediction method |
| CN109165774A (en) * | 2018-08-07 | 2019-01-08 | 燕山大学 | A kind of short-term photovoltaic power prediction technique |
| CN109272156A (en) * | 2018-09-12 | 2019-01-25 | 河海大学 | A kind of super short-period wind power probability forecasting method |
| CN109886461A (en) * | 2019-01-18 | 2019-06-14 | 昆仑(重庆)河湖生态研究院(有限合伙) | A kind of Runoff Forecast method and device |
| CN111275256A (en) * | 2020-01-18 | 2020-06-12 | 杭州电子科技大学 | Photovoltaic power generation power day-ahead prediction method based on image feature extraction |
| CN111444320A (en) * | 2020-06-16 | 2020-07-24 | 太平金融科技服务（上海）有限公司 | Text retrieval method and device, computer equipment and storage medium |
| CN112001554A (en) * | 2020-08-26 | 2020-11-27 | 山东德佑电气股份有限公司 | Short-term load prediction method based on parameter self-adaptive similar daily method |
| CN112507793A (en) * | 2020-11-05 | 2021-03-16 | 上海电力大学 | Ultra-short-term photovoltaic power prediction method |
| CN113222289A (en) * | 2021-06-02 | 2021-08-06 | 江南大学 | Energy power prediction method based on data processing |
| US20220197233A1 (en) * | 2020-12-18 | 2022-06-23 | Wuhan University | Wind power prediction method and system for optimizing deep transformer network |
| CN114970952A (en) * | 2022-04-14 | 2022-08-30 | 楚能新能源股份有限公司 | Photovoltaic output short-term prediction method and system considering environmental factors |
| WO2023004838A1 (en) * | 2021-07-26 | 2023-02-02 | 大连理工大学 | Wind power output interval prediction method |
| WO2023015460A1 (en) * | 2021-08-10 | 2023-02-16 | 中国电力科学研究院有限公司 | Wind process identification-based integrated optimization method and apparatus for wind power prediction |
| CN115860215A (en) * | 2022-11-29 | 2023-03-28 | 国网甘肃省电力公司电力科学研究院 | Photovoltaic and wind power generation power prediction method and system |
| US20230169230A1 (en) * | 2021-12-01 | 2023-06-01 | Central South University | Probabilistic wind speed forecasting method and system based on multi-scale information |
| CN116316591A (en) * | 2023-03-17 | 2023-06-23 | 山东科技大学 | Short-term photovoltaic power prediction method and system based on hybrid bidirectional gating cycle |
| CN116388174A (en) * | 2023-04-12 | 2023-07-04 | 南京工程学院 | Distributed photovoltaic cluster short-term output prediction method suitable for hierarchical scheduling |
| CN116432830A (en) * | 2023-03-21 | 2023-07-14 | 国网内蒙古东部电力有限公司呼伦贝尔供电公司 | Photovoltaic power prediction method, related computer equipment and storage medium |
| CN116454874A (en) * | 2023-04-12 | 2023-07-18 | 中国南方电网有限责任公司 | Wind power prediction method, wind power prediction device and electronic device |
-
2023
- 2023-09-12 CN CN202311166833.0A patent/CN116911467A/en active Pending
Patent Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105373857A (en) * | 2015-11-27 | 2016-03-02 | 许昌许继软件技术有限公司 | Photovoltaic power station irradiance prediction method |
| CN109165774A (en) * | 2018-08-07 | 2019-01-08 | 燕山大学 | A kind of short-term photovoltaic power prediction technique |
| CN109272156A (en) * | 2018-09-12 | 2019-01-25 | 河海大学 | A kind of super short-period wind power probability forecasting method |
| CN109886461A (en) * | 2019-01-18 | 2019-06-14 | 昆仑(重庆)河湖生态研究院(有限合伙) | A kind of Runoff Forecast method and device |
| CN111275256A (en) * | 2020-01-18 | 2020-06-12 | 杭州电子科技大学 | Photovoltaic power generation power day-ahead prediction method based on image feature extraction |
| CN111444320A (en) * | 2020-06-16 | 2020-07-24 | 太平金融科技服务（上海）有限公司 | Text retrieval method and device, computer equipment and storage medium |
| CN112001554A (en) * | 2020-08-26 | 2020-11-27 | 山东德佑电气股份有限公司 | Short-term load prediction method based on parameter self-adaptive similar daily method |
| CN112507793A (en) * | 2020-11-05 | 2021-03-16 | 上海电力大学 | Ultra-short-term photovoltaic power prediction method |
| US20220197233A1 (en) * | 2020-12-18 | 2022-06-23 | Wuhan University | Wind power prediction method and system for optimizing deep transformer network |
| CN113222289A (en) * | 2021-06-02 | 2021-08-06 | 江南大学 | Energy power prediction method based on data processing |
| WO2023004838A1 (en) * | 2021-07-26 | 2023-02-02 | 大连理工大学 | Wind power output interval prediction method |
| WO2023015460A1 (en) * | 2021-08-10 | 2023-02-16 | 中国电力科学研究院有限公司 | Wind process identification-based integrated optimization method and apparatus for wind power prediction |
| US20230169230A1 (en) * | 2021-12-01 | 2023-06-01 | Central South University | Probabilistic wind speed forecasting method and system based on multi-scale information |
| CN114970952A (en) * | 2022-04-14 | 2022-08-30 | 楚能新能源股份有限公司 | Photovoltaic output short-term prediction method and system considering environmental factors |
| CN115860215A (en) * | 2022-11-29 | 2023-03-28 | 国网甘肃省电力公司电力科学研究院 | Photovoltaic and wind power generation power prediction method and system |
| CN116316591A (en) * | 2023-03-17 | 2023-06-23 | 山东科技大学 | Short-term photovoltaic power prediction method and system based on hybrid bidirectional gating cycle |
| CN116432830A (en) * | 2023-03-21 | 2023-07-14 | 国网内蒙古东部电力有限公司呼伦贝尔供电公司 | Photovoltaic power prediction method, related computer equipment and storage medium |
| CN116388174A (en) * | 2023-04-12 | 2023-07-04 | 南京工程学院 | Distributed photovoltaic cluster short-term output prediction method suitable for hierarchical scheduling |
| CN116454874A (en) * | 2023-04-12 | 2023-07-18 | 中国南方电网有限责任公司 | Wind power prediction method, wind power prediction device and electronic device |
Non-Patent Citations (6)
| Title |
|---|
| ZULFIQAR AHMAD KHAN: "Towards efficient and effective renewable energy prediction via deep learning", 《ENERGY REPORTS》, vol. 8, pages 10230 - 10243 * |
| 国网湖南省电力有限公司编写组: "《湖南电力供需现状及预测》", 30 November 2021, 重庆大学出版社, pages: 151 - 152 * |
| 安鹏跃;孙?;: "基于相似日和回声状态网络的光伏发电功率预测", 智慧电力, no. 08, pages 38 - 43 * |
| 杜义林: "《物理实验与测试技术》", 31 January 2013, 中国科学技术大学出版社, pages: 172 - 174 * |
| 谈明高 等: "《光伏水泵运行特性与优化设计》", 30 April 2021, 江苏大学出版社, pages: 18 - 19 * |
| 谢少华 等: "基于 神经网络的光伏短期功率预测", 《浙江工业大学学报》, vol. 50, no. 6, pages 628 - 633 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113657662B (en) | Downscaling wind power prediction method based on data fusion | |
| CN110287540B (en) | Photovoltaic cell parameter identification method based on elite crowd differential evolution algorithm | |
| CN114462718A (en) | CNN-GRU wind power prediction method based on time sliding window | |
| CN113962433B (en) | Wind power prediction method and system integrating causal convolution and separable time convolution | |
| CN109242200B (en) | Wind power interval prediction method of Bayesian network prediction model | |
| CN110782071A (en) | Method for predicting wind power by convolutional neural network based on time-space characteristic fusion | |
| CN114676923A (en) | Method and device for predicting generated power, computer equipment and storage medium | |
| CN116128150A (en) | Photovoltaic intelligent output prediction method based on two-stage optimization | |
| CN111323847B (en) | Method and apparatus for determining weight ratios for analog integration algorithms | |
| CN115860215A (en) | Photovoltaic and wind power generation power prediction method and system | |
| CN115438841A (en) | Training method and prediction method based on artificial intelligence accurate prediction rainfall model | |
| CN114882373A (en) | Multi-feature fusion sandstorm prediction method based on deep neural network | |
| CN112418504B (en) | Wind speed prediction method based on mixed variable selection optimization deep belief network | |
| CN116911467A (en) | Renewable energy output prediction method, device and storage medium | |
| CN116502074A (en) | Model fusion-based photovoltaic power generation power prediction method and system | |
| CN116050592A (en) | Multi-dimensional photovoltaic power prediction method and system | |
| CN115587664A (en) | Solar radiation prediction method and system based on graph convolution long-term and short-term memory network | |
| CN115173421A (en) | Probability optimal power flow calculation method based on progressive Latin hypercube sampling | |
| CN114971007A (en) | Photovoltaic power generation prediction method and system based on multi-scale graph convolutional neural network | |
| CN113962432A (en) | Wind power prediction method and system integrating three-dimensional convolution and light-weight convolution threshold unit | |
| CN117151303B (en) | Ultra-short-term solar irradiance prediction method and system based on hybrid model | |
| CN116227743B (en) | Photovoltaic power generation power abnormal rate determining method and system based on tuna swarm algorithm | |
| CN117933363B (en) | Method and system for predicting short-term wind power of newly-built wind power plant based on sample migration | |
| CN117332905A (en) | Photovoltaic power prediction method and related device | |
| CN117350167A (en) | Equivalent method, system, storage medium and equipment for wind power plant |
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 |