CN113537582B - Photovoltaic power ultra-short-term prediction method based on short-wave radiation correction - Google Patents
Photovoltaic power ultra-short-term prediction method based on short-wave radiation correction Download PDFInfo
- Publication number
- CN113537582B CN113537582B CN202110746353.6A CN202110746353A CN113537582B CN 113537582 B CN113537582 B CN 113537582B CN 202110746353 A CN202110746353 A CN 202110746353A CN 113537582 B CN113537582 B CN 113537582B
- Authority
- CN
- China
- Prior art keywords
- output
- short
- wave radiation
- data
- photovoltaic power
- 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.)
- Active
Links
- 230000005855 radiation Effects 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000012937 correction Methods 0.000 title claims abstract description 15
- 238000004364 calculation method Methods 0.000 claims abstract description 14
- 238000004088 simulation Methods 0.000 claims abstract description 10
- 210000004027 cell Anatomy 0.000 claims description 39
- 239000013598 vector Substances 0.000 claims description 17
- 239000011159 matrix material Substances 0.000 claims description 15
- 210000002569 neuron Anatomy 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 230000007613 environmental effect Effects 0.000 claims description 3
- 210000002364 input neuron Anatomy 0.000 claims description 3
- 230000001537 neural effect Effects 0.000 claims description 3
- 238000001556 precipitation Methods 0.000 claims description 3
- 238000003064 k means clustering Methods 0.000 claims description 2
- 238000000638 solvent extraction Methods 0.000 claims description 2
- 230000007774 longterm Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 10
- 238000010248 power generation Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Marketing (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Probability & Statistics with Applications (AREA)
- Public Health (AREA)
- Evolutionary Biology (AREA)
- Primary Health Care (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Water Supply & Treatment (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a photovoltaic power station power ultra-short term prediction method based on short wave radiation correction, which comprises the steps of extracting numerical weather forecast (NWP) of a photovoltaic power station, correcting short wave radiation by using a long-term and short-term network (LSTM), predicting an Extreme Learning Machine (ELM), carrying out simulation calculation and evaluating indexes.
Description
Technical Field
The invention relates to the technical field of photovoltaic power stations, in particular to a photovoltaic power station power ultra-short-term prediction method based on short-wave radiation correction.
Background
The photovoltaic power generation has stronger day and night periodicity and seasonal periodicity, and is a typical intermittent power supply; meanwhile, the solar energy collector is very sensitive to the surface solar radiation intensity, has strong randomness of output, and has great influence on frequency modulation, peak regulation, standby and the like of a power grid. The photovoltaic power prediction is a basic key technology for improving the control and scheduling performance of a photovoltaic power station and ensuring the safe and stable operation of the high-ratio photovoltaic power generation access power grid.
The photovoltaic power ultra-short-term prediction refers to prediction and forecast of 4 hours in the future from the prediction moment, and the time resolution is 15 minutes. The significance of the day-ahead prediction is that ultra-short-term photovoltaic power prediction can provide power transient information.
The existing photovoltaic power ultra-short-term prediction generally establishes a mapping relation between historical input data and future power output, and can directly predict a future power value according to historical NWP data, so that higher prediction accuracy is obtained. For the artificial intelligence method, the method has great advantages in processing the nonlinear time series, but cannot reflect the dynamic characteristics of the system. Overall, existing predictions cannot track future power trends.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide the photovoltaic power ultra-short-term prediction method based on the short-wave radiation correction, which has clear physical significance, considers rich time-space correlation among data, and has the advantages of scientific and reasonable method, strong applicability, high prediction precision and accurate prediction result.
The technical scheme adopted for realizing the aim of the invention is as follows: a photovoltaic power ultra-short term prediction method based on short wave radiation correction is characterized by comprising the following steps: it comprises the following steps:
1) partitioning of weather types based on Elkan-K-means clustering
Numerical Weather forecast (NWP) information includes 10 meter wind speed, 30 meter wind speed, 50 meter wind speed, 70 meter wind speed, 10 meter wind direction, 30 meter wind direction, 50 meter wind direction, 70 meter wind direction, temperature, 2 meter relative humidity, 2m humidity, long wave radiation, short wave radiation, cloud cover, air pressure, precipitation;
in order to avoid the influence of redundant information on the prediction precision, a pearson correlation coefficient is utilized to preliminarily screen weather forecast meteorological factors and corresponding photovoltaic power, the pearson correlation coefficient of each meteorological factor sequence and the photovoltaic power sequence is respectively solved, a factor showing a positive correlation relation with the photovoltaic power is selected, and short-wave radiation with the maximum person coefficient is selected as the relevant information input;
2) short wave radiation correction using long-short term memory network (LSTM)
Based on the relevant information extracted in the step 1), judging that when short-wave radiation is predicted, under certain specific environmental conditions, the rule that the predicted value is lower than the actual value is utilized, and a long-short term memory network (LSTM) is adopted to deeply mine complex characteristic information in sample data by utilizing the characteristics of a training sample to correct the data;
the internal calculation process of the long-short term memory network (LSTM) neural unit is as follows:
the method comprises the following steps that a Sigmoid layer of a forgetting gate determines what information is discarded from a cell state, the cell state comprises basic attribute characteristics of photovoltaic power station weather forecast (NWP), therefore, future short wave radiation can be corrected, when new operation data are input, people hope to forget to compare old numerical weather forecast (NWP) distribution information and discard old information from cells, the step of determining is completed through the forgetting gate layer, and a specific function expression of calculation of the forgetting gate is as follows:
ft＝sigmoid(θf·[ht-1,xt]+bf) (1)
the next step to decide what information to store in the cell state is to take as input the shortwave radiation provided by the NWP, 2m humidity, temperature, 10 m wind speed, 70 m wind speed, 30m wind direction, the Sigmoid layer called "input gate layer" decides which values to update, expressed by equation (2), the next tanh layer creates a candidate vector Ct, which will be added to the state of the cell, expressed by equation (3), and in the next step, combining equations (2) and (3) to create the updated values, expressed by equation (4),
it＝sigmoid(Wi·[ht-1,xt]+bi) (2)
renewal of cell status
This output will be based on the cell state, but will be a filtered version, first running a Sigmoid layer to determine the fraction of the cell state to output, then passing the cell state through tanh, normalizing the value to between-1 and 1, and multiplying it by the output of the Sigmoid gate, two steps to determine the storage of the input in the cell state, expressed by equations (5), 6,
ot＝sigmoid(Wo·[ht-1,xt]+bo) (5)
ht＝ot·tanh(Ct) (6)
in the formula itIs the output of the input gate or gates,
as a candidate for the current layer, C may be added to the cell statetIs the current memory cell state, and the whole process is the process of updating the memory cell state at the previous moment, namely discarding useless information and adding new information, otIs the output of the output gate, htThe final output of the LSTM at the current moment is output, and the corrected short wave radiation at the corresponding time interval is output;
3) training of Extreme learning machine (Extreme learning machine)
If the learning data sample of the ELM is { xj, tj }, and xj and tj respectively represent input data and output data of the model, wherein the input data are the short-wave radiation and the historical photovoltaic power corrected in the step 2), and the output data are that the ultra-short-term photovoltaic power predictions all belong to a real number set R, the mathematical expression of the single hidden layer neural network is as follows:
in the formula: kcon is the number of hidden layer nodes; beta is aiIs the output weight; g is an activation function; omegaiThe connection weight between the input neuron and the ith hidden neuron; omegai·xjThe inner product of the connection weight and the data sample; biBias compensation values for the ith hidden layer neuron; t is tjIs an output vector; m is the total number of samples;
setting T as a matrix form of an output vector, and beta as an output weight vector, and obtaining an expression of the output vector, namely, T is in an H beta form, wherein H is a hidden layer output matrix of the ELM neural network;
h is in a matrix form
The training process of the ELM model comprises the following steps: firstly, determining the number of neurons, and randomly distributing node parameters; then, calculating an output matrix of the Mth row and the Kcon column of the hidden layer, and calculating the photovoltaic output power of the output matrix;
4) simulation calculation
Simulation input quantity: analyzing the actually measured data of the photovoltaic power station; inputting historical data: inputting numerical weather forecast (NWP) data of predicted months for the first two monthly history powers of each quarter; the data sampling interval is 15min, and a final power ultra-short-term prediction result is obtained according to the steps 1) to 3);
5) evaluation index
Let PmiIs the actual average power, P, of the i periodpiPredicted power for i period, CiFor the total starting capacity in the period i, the average absolute error is defined as formula (9) if n is the number of all samples:
the root mean square error is defined by equation (11):
and 4), inputting simulation input quantity, and carrying out error calculation on the predicted power calculated by the model and the actual measured power through the average absolute error (9) and the root-mean-square error (10) in the step 5) to obtain the prediction accuracy.
The invention relates to a photovoltaic power station power ultra-short term prediction method based on short wave radiation correction, which comprises the steps of extracting numerical weather forecast (NWP) of a photovoltaic power station, correcting short wave radiation by utilizing a long-term and short-term network (LSTM), predicting Extreme Learning Machine (ELM), simulating calculation and evaluating indexes.
Drawings
FIG. 1 is a block diagram of a photovoltaic power station power ultra-short term prediction method based on short wave radiation correction according to the present invention;
FIG. 2 is a schematic diagram of a numerical weather forecast (NWP) selection of photovoltaic power at a certain time;
FIG. 3 is a schematic diagram showing comparison between predicted values and actual values in spring for a certain day;
FIG. 4 is a schematic diagram showing comparison between predicted values and actual values of a day in summer;
FIG. 5 is a schematic diagram showing the comparison between the predicted value and the actual value of a day in autumn;
fig. 6 is a schematic diagram showing comparison between predicted values and actual values of a certain day in winter.
Detailed Description
The method for ultra-short-term power prediction of a photovoltaic power station based on short-wave radiation correction is further described below with reference to the accompanying drawings and specific embodiments.
With reference to fig. 1 and fig. 2, the ultrashort-term photovoltaic power station power prediction method based on short-wave radiation correction of the present invention includes the following steps:
1) numerical weather forecast selection
Numerical Weather Prediction (NWP) information includes 10 meter wind speed, 30 meter wind speed, 50 meter wind speed, 70 meter wind speed, 10 meter wind direction, 30 meter wind direction, 50 meter wind direction, 70 meter wind direction, temperature, 2 meter relative humidity, 2m humidity, long wave radiation, short wave radiation, cloud cover, air pressure, precipitation, and the like.
In order to avoid the influence of redundant information on the prediction precision, the pearson correlation coefficient is utilized to preliminarily screen weather forecast meteorological factors and corresponding photovoltaic power, and the pearson correlation coefficient of each meteorological factor sequence and the corresponding photovoltaic power sequence is respectively obtained. And selecting a factor which has positive correlation with the photovoltaic power, and selecting the short-wave radiation with the maximum person coefficient as the relevant information input.
2) Short-wave radiation correction using long-short term memory network (LSTM)
Based on the related information extracted in step 1), it can be determined that, when short-wave radiation is predicted, the prediction deviation has certain regularity under certain specific environmental conditions, such as sunshine time in the daytime, and if most of the predicted values are lower than the actual values, the data can be corrected by using the regularity. Because the correction cannot be expressed by a specific function, a long-short term memory network (LSTM) is a popular deep learning method at present, and has the advantages that the characteristics of the training sample can be fully utilized, the complex characteristic information in the sample data can be deeply mined, and the high-dimensional function approximation can be completed.
The internal calculation process for the LSTM neural unit is as follows:
the Sigmoid layer of the forgetting gate determines what information is discarded from the cell state, the cell state comprises basic attribute characteristics of photovoltaic power station weather forecast (NWP), therefore, the future short wave radiation can be corrected, when new operation data is input, people hope to forget to compare old numerical weather forecast (NWP) distribution information, and discard the old information from the cell, and the step of determining is completed through the forgetting gate layer. The specific function expression of the forgetting gate calculation method is
ft＝sigmoid(θf·[ht-1,xt]+bf) (1)
The next step to decide what information is stored in the cell state is to take as input the shortwave radiation provided by the NWP, 2 meters humidity, temperature, 10 meters wind speed, 70 meters wind speed, 30 meters wind direction, the Sigmoid layer called "input gate layer" decides which values to update represented by equation (2), the next tanh layer creates a candidate vector Ct, which is to be added to the state of the cell represented by equation (3), and in the next step, the two vectors are combined to create an updated value represented by equation (4),
it＝sigmoid(Wi·[ht-1,xt]+bi) (2)
renewal of cell status
This output will be based on the cell state, but will be a filtered version. First, a Sigmoid layer is run to determine the fraction of cell states to be output, then the cell states are normalized to a value between-1 and 1 by tanh and multiplied by the output of the Sigmoid gate, two steps of determining to store inputs in the cell states are expressed by equations (5) and (6),
ot＝sigmoid(Wo·[ht-1,xt]+bo) (5)
ht＝ot·tanh(Ct) (6)
in the formula itIs the output of the input gate or gates,
as a candidate for the current layer, C may be added to the cell statetIs the current memory cell state, and the whole process is the process of updating the memory cell state at the previous moment, namely discarding useless information and adding new information, otIs the output of the output gate, htThe final output of the LSTM at the current moment is output, and the corrected short wave radiation at the corresponding time interval is output;
(3) training of Extreme learning machine (Extreme learning machine)
If the learning data sample of the ELM is { xj, tj }, and xj and tj respectively represent input data and output data of the model, wherein the input data are the short-wave radiation and the historical photovoltaic power corrected in the step 2), and the output data are that the ultra-short-term photovoltaic power predictions all belong to a real number set R, the mathematical expression of the single hidden layer neural network is
In the formula: kcon is the number of hidden layer nodes; beta is aiIs the output weight; g is an activation function; omegaiThe connection weight between the input neuron and the ith hidden neuron; omegai·xjThe inner product of the connection weight and the data sample; biBias compensation values for the ith hidden layer neuron; t is tjIs an output vector; m is the total number of samples.
Let T be the matrix form of the output vector and β be the output weight vector, an expression of the output vector can be obtained, i.e., T ═ H β form, where H is the hidden layer output matrix of the ELM neural network.
H is in a matrix form
The training process of the ELM model comprises the following steps: firstly, determining the number of neurons, and randomly distributing node parameters; then, the output matrix of the Mth row and the Kcon column of the hidden layer is calculated, and the photovoltaic output power of the output matrix is calculated.
4) Simulation calculation
Simulation input quantity: analyzing the actually measured data of the photovoltaic power station; inputting historical data: inputting numerical weather forecast (NWP) data of predicted months for the first two monthly history powers of each quarter; the data sampling interval is 15min, and a final power ultra-short-term prediction result is obtained according to the steps 1) to 3);
5) evaluation index
Let PmiIs the actual average power of the i period, PpiPredicted power for i period, CiFor the total starting capacity in the period i, the average absolute error is defined as formula (9) if n is the number of all samples:
the root mean square error is defined by equation (11):
and 4), inputting simulation input quantity, and carrying out error calculation on the predicted power calculated by the model and the actual measured power through the average absolute error (9) and the root-mean-square error (10) in the step 5) to obtain the prediction accuracy.
Detailed description of the invention
The method takes measured data and numerical weather forecast (NWP) data of a certain photovoltaic power station as examples for analysis, and the sampling interval is 15 min. The installed capacity of the photovoltaic power station is 30 MW; the results obtained by prediction with different prediction models are shown in fig. 3-6, and fig. 3 is a comparison diagram of the predicted value and the true value in a certain day in spring; FIG. 4 is a schematic diagram showing comparison between predicted values and actual values of a day in summer; FIG. 5 is a schematic diagram showing the comparison between the predicted value and the actual value of a day in autumn; fig. 6 is a schematic diagram showing comparison between a predicted value and a true value in a certain day in winter, and specific evaluation indexes are shown in table 1, so that it can be seen that the prediction effect of the used LSTM-ELM prediction model is superior to that of other prediction methods.
TABLE 1 comparison of evaluation indices of prediction models
Table 1 Comparison of evaluation indexes of prediction model
The description of the present invention is not intended to be exhaustive or to limit the scope of the claims, and those skilled in the art will be able to conceive of other substantially equivalent alternatives, without inventive step, based on the teachings of the embodiments of the present invention, within the scope of the present invention.
Claims (1)
1. A photovoltaic power ultra-short term prediction method based on short wave radiation correction is characterized by comprising the following steps: it comprises the following steps:
1) partitioning of weather types based on Elkan-K-means clustering
Numerical Weather forecast (NWP) information includes 10 meter wind speed, 30 meter wind speed, 50 meter wind speed, 70 meter wind speed, 10 meter wind direction, 30 meter wind direction, 50 meter wind direction, 70 meter wind direction, temperature, 2 meter relative humidity, 2m humidity, long wave radiation, short wave radiation, cloud cover, air pressure, precipitation;
in order to avoid the influence of redundant information on the prediction precision, a pearson correlation coefficient is utilized to preliminarily screen weather forecast meteorological factors and corresponding photovoltaic power, the pearson correlation coefficient of each meteorological factor sequence and the photovoltaic power sequence is respectively solved, a factor showing a positive correlation relation with the photovoltaic power is selected, and short-wave radiation with the maximum person coefficient is selected as the relevant information input;
2) short-wave radiation correction using long-short term memory network (LSTM)
Based on the relevant information extracted in the step 1), judging that when short-wave radiation is predicted, under certain specific environmental conditions, the rule that the predicted value is lower than the actual value is utilized, and a long-short term memory network (LSTM) is adopted to deeply mine complex characteristic information in sample data by utilizing the characteristics of a training sample to correct the data;
the internal calculation process of the long-short term memory network (LSTM) neural unit is as follows:
the method comprises the following steps that a Sigmoid layer of a forgetting gate determines what information is discarded from a cell state, the cell state comprises basic attribute characteristics of photovoltaic power station weather forecast (NWP), therefore, future short wave radiation can be corrected, when new operation data are input, people hope to forget to compare old numerical weather forecast (NWP) distribution information and discard old information from cells, the step of determining is completed through the forgetting gate layer, and a specific function expression of calculation of the forgetting gate is as follows:
ft＝sigmoid(θf·[ht-1,xt]+bf) (1)
next, it is decided what information to store in the cell state, first, which values to update are decided by the Sigmoid layer called "input gate layer" using as input the shortwave radiation provided by NWP, 2m humidity, temperature, 10 m wind speed, 70 m wind speed, 30m wind direction, expressed by equation (2), next, a candidate vector Ct is created by a tanh layer, which candidate vector Ct will be added to the state of the cell and expressed by equation (3), and in the next step, the updated values are created by combining equations (2) and (3) and expressed by equation (4),
it＝sigmoid(Wi·[ht-1,xt]+bi) (2)
renewal of cell status
Ct＝it⊙Ct+ft⊙Ct-1 (4)
This output will be based on the cell state, but will be a filtered version, first running a Sigmoid layer to determine the fraction of the cell state to output, then passing the cell state through tanh, normalizing the value to between-1 and 1, and multiplying it by the output of the Sigmoid gate, two steps to determine the storage of the input in the cell state, expressed by equations (5), 6,
ot＝sigmoid(Wo·[ht-1,xt]+bo) (5)
ht＝ot·tanh(Ct) (6)
in the formula itIs the output of the input gate or gates,
as a candidate for the current layer, C may be added to the cell statetIs the current memory cell state, and the whole process is the process of updating the memory cell state at the previous moment, namely discarding useless information and adding new information, otIs the output of the output gate, htThe final output of the LSTM at the current moment is output, and the corrected short wave radiation at the corresponding time interval is output;
3) training of Extreme learning machine (Extreme learning machine)
If the learning data sample of the ELM is { xj, tj }, and xj and tj respectively represent input data and output data of the model, wherein the input data are the short-wave radiation and the historical photovoltaic power corrected in the step 2), and the output data are that ultra-short-term photovoltaic power predictions all belong to a real number set R, the mathematical expression of the single hidden layer neural network is as follows:
in the formula: kcon is the number of hidden layer nodes; beta is aiIs the output weight; g is an activation function; omegaiThe connection weight between the input neuron and the ith hidden neuron; omegai·xjThe inner product of the connection weight and the data sample; biBias compensation values for the ith hidden layer neuron; t is tjIs an output vector; m is the total number of samples;
setting T as a matrix form of an output vector, and setting beta as an output weight vector, and obtaining an expression of the output vector, namely a form of T ═ H beta, wherein H is a hidden layer output matrix of the ELM neural network;
h is in a matrix form
The training process of the ELM model comprises the following steps: firstly, determining the number of neurons, and randomly distributing node parameters; then, calculating an output matrix of the Mth row and the Kcon column of the hidden layer, and calculating the photovoltaic output power of the output matrix;
4) simulation calculation
Simulation input quantity: analyzing the actually measured data of the photovoltaic power station; inputting historical data: inputting numerical weather forecast (NWP) data of predicted months for the first two monthly history powers of each quarter; the data sampling interval is 15min, and a final power ultra-short-term prediction result is obtained according to the steps 1) to 3);
5) evaluation index
Let PmiIs the actual average power of the i period, PpiPredicted power for i period, CiFor the total starting capacity in the period i, the average absolute error is defined as formula (9) if n is the number of all samples:
the root mean square error is defined by equation (11):
and 4), inputting simulation input quantity, and carrying out error calculation on the predicted power calculated by the model and the actual measured power through the average absolute error (9) and the root-mean-square error (10) in the step 5) to obtain the prediction accuracy.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110746353.6A CN113537582B (en) | 2021-07-02 | 2021-07-02 | Photovoltaic power ultra-short-term prediction method based on short-wave radiation correction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110746353.6A CN113537582B (en) | 2021-07-02 | 2021-07-02 | Photovoltaic power ultra-short-term prediction method based on short-wave radiation correction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113537582A CN113537582A (en) | 2021-10-22 |
| CN113537582B true CN113537582B (en) | 2022-05-24 |
Family
ID=78097542
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110746353.6A Active CN113537582B (en) | 2021-07-02 | 2021-07-02 | Photovoltaic power ultra-short-term prediction method based on short-wave radiation correction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113537582B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114006369B (en) * | 2021-10-28 | 2022-08-30 | 华北电力大学 | Regional wind and light station power joint prediction method and device, electronic equipment and storage medium |
| CN116245259B (en) * | 2023-05-11 | 2023-10-31 | 华能山东泰丰新能源有限公司 | Photovoltaic power generation prediction method and device based on depth feature selection and electronic equipment |
| CN116757101B (en) * | 2023-08-21 | 2023-11-07 | 湖南科技大学 | Cabin wind speed correction method and system based on mechanism model and neural network |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108280551A (en) * | 2018-02-02 | 2018-07-13 | 华北电力大学 | A kind of photovoltaic power generation power prediction method using shot and long term memory network |
| CN108985965A (en) * | 2018-06-22 | 2018-12-11 | 华北电力大学 | A kind of photovoltaic power interval prediction method of combination neural network and parameter Estimation |
| CN109284870A (en) * | 2018-10-08 | 2019-01-29 | 南昌大学 | Short-term method for forecasting photovoltaic power generation quantity based on shot and long term Memory Neural Networks |
| CN109787295A (en) * | 2019-02-26 | 2019-05-21 | 东北电力大学 | It is a kind of meter and wind power plant state ultrashort-term wind power predict calculation method |
| CN110363360A (en) * | 2019-07-24 | 2019-10-22 | 广东工业大学 | A kind of short-term wind power forecast method, device and equipment |
| CN110738363A (en) * | 2019-09-30 | 2020-01-31 | 华中科技大学 | photovoltaic power generation power prediction model and construction method and application thereof |
| CN111008728A (en) * | 2019-11-01 | 2020-04-14 | 深圳供电局有限公司 | Method for predicting short-term output of distributed photovoltaic power generation system |
| CN111222674A (en) * | 2019-10-08 | 2020-06-02 | 南昌大学 | Short-term photovoltaic power generation capacity prediction method based on long-term and short-term memory neural network |
| CN111832825A (en) * | 2020-07-14 | 2020-10-27 | 山东大学 | Wind power prediction method and system integrating long-term and short-term memory network and extreme learning machine |
| CN112348271A (en) * | 2020-11-12 | 2021-02-09 | 华北电力大学 | Short-term photovoltaic power prediction method based on VMD-IPSO-GRU |
| CN112508245A (en) * | 2020-11-26 | 2021-03-16 | 郑州迈拓信息技术有限公司 | Method and system for predicting power generation capacity based on artificial intelligence photovoltaic cell panel glass breakage |
| CN112668806A (en) * | 2021-01-17 | 2021-04-16 | 中国南方电网有限责任公司 | Photovoltaic power ultra-short-term prediction method based on improved random forest |
| CN112906935A (en) * | 2020-12-21 | 2021-06-04 | 湘潭大学 | Method for predicting ultra-short-term power of wind power plant |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA3098670A1 (en) * | 2018-05-06 | 2019-11-14 | Strong Force TX Portfolio 2018, LLC | Methods and systems for improving machines and systems that automate execution of distributed ledger and other transactions in spot and forward markets for energy, compute, storage and other resources |
-
2021
- 2021-07-02 CN CN202110746353.6A patent/CN113537582B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108280551A (en) * | 2018-02-02 | 2018-07-13 | 华北电力大学 | A kind of photovoltaic power generation power prediction method using shot and long term memory network |
| CN108985965A (en) * | 2018-06-22 | 2018-12-11 | 华北电力大学 | A kind of photovoltaic power interval prediction method of combination neural network and parameter Estimation |
| CN109284870A (en) * | 2018-10-08 | 2019-01-29 | 南昌大学 | Short-term method for forecasting photovoltaic power generation quantity based on shot and long term Memory Neural Networks |
| CN109787295A (en) * | 2019-02-26 | 2019-05-21 | 东北电力大学 | It is a kind of meter and wind power plant state ultrashort-term wind power predict calculation method |
| CN110363360A (en) * | 2019-07-24 | 2019-10-22 | 广东工业大学 | A kind of short-term wind power forecast method, device and equipment |
| CN110738363A (en) * | 2019-09-30 | 2020-01-31 | 华中科技大学 | photovoltaic power generation power prediction model and construction method and application thereof |
| CN111222674A (en) * | 2019-10-08 | 2020-06-02 | 南昌大学 | Short-term photovoltaic power generation capacity prediction method based on long-term and short-term memory neural network |
| CN111008728A (en) * | 2019-11-01 | 2020-04-14 | 深圳供电局有限公司 | Method for predicting short-term output of distributed photovoltaic power generation system |
| CN111832825A (en) * | 2020-07-14 | 2020-10-27 | 山东大学 | Wind power prediction method and system integrating long-term and short-term memory network and extreme learning machine |
| CN112348271A (en) * | 2020-11-12 | 2021-02-09 | 华北电力大学 | Short-term photovoltaic power prediction method based on VMD-IPSO-GRU |
| CN112508245A (en) * | 2020-11-26 | 2021-03-16 | 郑州迈拓信息技术有限公司 | Method and system for predicting power generation capacity based on artificial intelligence photovoltaic cell panel glass breakage |
| CN112906935A (en) * | 2020-12-21 | 2021-06-04 | 湘潭大学 | Method for predicting ultra-short-term power of wind power plant |
| CN112668806A (en) * | 2021-01-17 | 2021-04-16 | 中国南方电网有限责任公司 | Photovoltaic power ultra-short-term prediction method based on improved random forest |
Non-Patent Citations (6)
| Title |
|---|
| A Two-Layer Nonlinear Combination Method for Short-Term Wind Speed Prediction Based on ELM, ENN, and LSTM;Min-Rong Chen 等;《IEEE Internet of Things Journal》;20190425;第6卷(第4期);第6997-7010页 * |
| 基于DTW-FCBF-LSTM模型的超短期风速预测;董治强;《电测与仪表》;20191216(第04期);第98-103页 * |
| 基于VMD和LSTM的超短期风速预测;王俊等;《电力***保护与控制》;20191216(第11期);第51-58页 * |
| 基于深度学习的LSTM光伏预测;崔承刚等;《上海电力学院学报》;20191215(第06期);第38-46页 * |
| 基于集合经验模态分解和套索算法的短期风速组合变权预测模型研究;杨磊等;《电力***保护与控制》;20200518(第10期);第87-96页 * |
| 考虑空间相关性采用LSTM神经网络的光伏出力短期预测方法;王志远等;《电力***及其自动化学报》;20190722(第05期);第83-90页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113537582A (en) | 2021-10-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109711620B (en) | Short-term power load prediction method based on GRU neural network and transfer learning | |
| Gao et al. | Day-ahead power forecasting in a large-scale photovoltaic plant based on weather classification using LSTM | |
| CN113537582B (en) | Photovoltaic power ultra-short-term prediction method based on short-wave radiation correction | |
| CN110766212B (en) | Ultra-short-term photovoltaic power prediction method for historical data missing electric field | |
| CN110110912B (en) | Photovoltaic power multi-model interval prediction method | |
| CN109948845A (en) | A kind of distribution network load shot and long term Memory Neural Networks prediction technique | |
| CN114792156B (en) | Photovoltaic output power prediction method and system based on curve characteristic index clustering | |
| CN111695736B (en) | Photovoltaic power generation short-term power prediction method based on multi-model fusion | |
| CN112036658B (en) | Wind power real-time prediction method based on forgetting long-term and short-term memory with difference | |
| CN111915092A (en) | Ultra-short-term wind power prediction method based on long-time and short-time memory neural network | |
| CN113516271A (en) | Wind power cluster power day-ahead prediction method based on space-time neural network | |
| CN115860177A (en) | Photovoltaic power generation power prediction method based on combined machine learning model and application thereof | |
| CN112163689A (en) | Short-term load quantile probability prediction method based on depth Attention-LSTM | |
| CN112329990A (en) | User power load prediction method based on LSTM-BP neural network | |
| CN115130741A (en) | Multi-model fusion based multi-factor power demand medium and short term prediction method | |
| CN115796004A (en) | Photovoltaic power station ultra-short term power intelligent prediction method based on SLSTM and MLSTNet models | |
| CN116526473A (en) | Particle swarm optimization LSTM-based electrothermal load prediction method | |
| CN115115125A (en) | Photovoltaic power interval probability prediction method based on deep learning fusion model | |
| CN117977568A (en) | Power load prediction method based on nested LSTM and quantile calculation | |
| CN116167531A (en) | Photovoltaic power generation prediction method based on digital twin | |
| CN114169639A (en) | Enterprise power consumption prediction method | |
| CN116341613A (en) | Ultra-short-term photovoltaic power prediction method based on Informar encoder and LSTM | |
| CN111178612B (en) | LSTM load prediction method of grid user based on big data ODPS engine | |
| CN116523142A (en) | Photovoltaic cluster interval prediction method considering space-time characteristic clustering | |
| CN113449847B (en) | Offshore wind power rolling prediction method considering second-level time series wind speed change |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |