CN115563848A - Distributed photovoltaic total radiation prediction method and system based on deep learning - Google Patents
Distributed photovoltaic total radiation prediction method and system based on deep learning Download PDFInfo
- Publication number
- CN115563848A CN115563848A CN202110741009.8A CN202110741009A CN115563848A CN 115563848 A CN115563848 A CN 115563848A CN 202110741009 A CN202110741009 A CN 202110741009A CN 115563848 A CN115563848 A CN 115563848A
- Authority
- CN
- China
- Prior art keywords
- total radiation
- data
- prediction
- distributed photovoltaic
- historical
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/10—Numerical modelling
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Geometry (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a distributed photovoltaic total radiation prediction method and a distributed photovoltaic total radiation prediction system based on deep learning, which comprise the following steps: screening out numerical weather forecast data in a prediction time period based on a predetermined meteorological influence factor of distributed photovoltaic total radiation; substituting the screened numerical weather forecast data of the prediction time interval into a pre-established total radiation prediction model based on deep learning to obtain a prediction value of the distributed photovoltaic total radiation of the prediction time interval; the weather influence factor of the distributed photovoltaic total radiation is determined by utilizing the correlation between the historical total radiation data of the weather monitoring station in the prediction region; and the total radiation prediction model carries out iterative training on the numerical mode prediction data and the total radiation monitoring data by utilizing a neural network model to obtain distributed photovoltaic total radiation prediction data corresponding to the prediction time period. According to the technical scheme provided by the invention, main meteorological influence factors influencing the distributed photovoltaic total radiation are fully considered, and the accuracy of the distributed photovoltaic total radiation prediction data is improved.
Description
Technical Field
The invention relates to the technical field of new energy, in particular to a distributed photovoltaic total radiation prediction method and system based on deep learning.
Background
With the development of new energy, the proportion of the distributed photovoltaic power generation in the power load is gradually increased. The distributed photovoltaic system has the advantages of large quantity, small scale, scattered geographic positions, complex operation environment, laggard communication infrastructure and unclear operation and maintenance subject. The main influence factor of distributed photovoltaic power generation is total radiation, but distributed photovoltaic total radiation data monitoring has the problems of low data acquisition rate, poor reliability, lack of control capability, insufficient technical support capability of an automatic master station and the like in information acquisition due to a plurality of complex reasons.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a distributed photovoltaic total radiation prediction method based on deep learning, which comprises the following steps:
screening out numerical weather forecast data in a prediction time period based on a predetermined meteorological influence factor of distributed photovoltaic total radiation;
substituting the screened numerical weather forecast data of the prediction time interval into a pre-established total radiation prediction model based on deep learning to obtain a prediction value of the distributed photovoltaic total radiation of the prediction time interval;
the meteorological influence factor of the distributed photovoltaic total radiation is determined by utilizing the correlation between the historical total radiation data of the meteorological monitoring station in the forecast area;
and the total radiation prediction model carries out iterative training on the numerical mode prediction data and the total radiation monitoring data by utilizing a neural network model to obtain distributed photovoltaic total radiation prediction data corresponding to the prediction time period.
Preferably, the determining process of the weather influence factor of the distributed photovoltaic total radiation includes:
determining a nearest weather monitoring station based on the position information of all distributed photovoltaic points in the area;
acquiring historical total radiation time sequence data of the weather station, and acquiring time sequence data of a plurality of weather elements corresponding to the historical total radiation time sequence data;
and determining correlation coefficients of the acquired time sequence data of the meteorological elements corresponding to the historical total radiation time sequence data and the historical total radiation time sequence data, and determining meteorological influence factors of the distributed photovoltaic total radiation based on the correlation coefficients.
Further, the calculation formula of the correlation coefficient between the acquired time series data of the plurality of meteorological elements corresponding to the historical total radiation time series data and the historical total radiation time series data is as follows:
in the formula, r x Correlation coefficient of x type meteorological element and historical total radiation monitoring data, x i Is the meteorological data of the ith time in the time series data of the x-th type meteorological elements, x is the average value of the meteorological data of each time in the time series data of the x-th type meteorological elements, y i The data is historical total radiation monitoring data at the ith moment, y is the average value of the historical total radiation monitoring data, and n is the total number of the moments.
Preferably, the weather influencing factor includes:
total radiation, air temperature, relative humidity, sea level air pressure and wind speed.
Preferably, the process of establishing the total radiation prediction model includes:
taking the screened numerical weather forecast data influencing the distributed photovoltaic total radiation in the historical time period as initial neural network model input data, taking total radiation monitoring data in the historical time period as initial neural network model output data, and determining the center and variance of a basis function of the initial neural network model according to the initial neural network model input data;
and performing iterative training on the weight coefficient of the initial neural network model based on the center and the variance of the basis function to obtain a deep learning total radiation prediction model.
Further, the iteratively training the weight coefficient of the initial neural network model based on the center and the variance of the basis function to obtain a deep learning total radiation prediction model includes:
setting an initial value of a weight coefficient of the initial neural network model;
performing iterative training on the weight coefficient of the initial neural network model based on the center, the variance and the initial value of the weight coefficient of the basis function, and determining the weight coefficient of the initial neural network during iterative convergence by using a least square method;
determining the corresponding relation between the total radiation monitoring data in the historical time period and the screened numerical weather forecast data which influences the distributed photovoltaic total radiation in the historical time period by using the weight coefficient of the initial neural network;
and determining a deep learning total radiation prediction model based on the corresponding relation between the total radiation monitoring data in the historical time period and the numerical weather forecast data which influence the distributed photovoltaic total radiation in the screened historical time period.
Further, the center of the basis function is determined by a self-organizing selection method based on a K-means clustering algorithm.
Further, the variance of the basis functions is determined based on the maximum distance between the centers of the basis functions.
The invention provides a distributed photovoltaic total radiation prediction system based on deep learning based on the same inventive concept, and the system comprises:
the screening module is used for screening out numerical weather forecast data in a prediction time period based on a predetermined meteorological influence factor of the distributed photovoltaic total radiation;
the prediction module is used for substituting the screened numerical weather forecast data of the prediction time interval into a pre-established total radiation prediction model based on deep learning to obtain a prediction value of the distributed photovoltaic total radiation of the prediction time interval;
the meteorological influence factor of the distributed photovoltaic total radiation is determined by utilizing the correlation between the historical total radiation data of the meteorological monitoring station in the forecast area;
and the total radiation prediction model carries out iterative training on the numerical mode prediction data and the total radiation monitoring data by utilizing a neural network model to obtain distributed photovoltaic total radiation prediction data corresponding to a prediction time period.
Preferably, the determining process of the weather influence factor of the distributed photovoltaic total radiation includes:
determining a nearest weather monitoring station based on the position information of all distributed photovoltaic points in the area;
acquiring historical total radiation time sequence data of the weather station, and acquiring time sequence data of a plurality of weather elements corresponding to the historical total radiation time sequence data;
and determining correlation coefficients of the acquired time sequence data of the meteorological elements corresponding to the historical total radiation time sequence data and the historical total radiation time sequence data, and determining meteorological influence factors of the distributed photovoltaic total radiation based on the correlation coefficients.
Compared with the closest prior art, the invention has the following beneficial effects:
the invention provides a distributed photovoltaic total radiation prediction method and a distributed photovoltaic total radiation prediction system based on deep learning, which comprise the following steps: screening out numerical weather forecast data in a prediction time period based on a predetermined meteorological influence factor of distributed photovoltaic total radiation; substituting the screened numerical weather forecast data of the prediction time interval into a pre-established total radiation prediction model based on deep learning to obtain a prediction value of the distributed photovoltaic total radiation of the prediction time interval; the weather influence factor of the distributed photovoltaic total radiation is determined by utilizing the correlation between the historical total radiation data of the weather monitoring station in the prediction region; and the total radiation prediction model carries out iterative training on the numerical mode prediction data and the total radiation monitoring data by utilizing a neural network model to obtain distributed photovoltaic total radiation prediction data corresponding to a prediction time period. According to the technical scheme provided by the invention, main meteorological influence factors influencing the distributed photovoltaic total radiation are fully considered, and the accuracy of the distributed photovoltaic total radiation prediction data is improved.
Drawings
FIG. 1 is a flow chart of a distributed photovoltaic total radiation prediction method based on deep learning according to the present invention;
FIG. 2 is a schematic diagram of a deep learning-based total radiation prediction model provided by the present invention;
fig. 3 is a structural diagram of a distributed photovoltaic total radiation prediction system based on deep learning according to the present invention.
Detailed Description
The following describes embodiments of the present invention in further detail with reference to the accompanying drawings.
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
The invention provides a distributed photovoltaic total radiation prediction method based on deep learning, and aims to predict distributed photovoltaic total radiation based on a predetermined meteorological influence factor of the distributed photovoltaic total radiation, so that the accuracy of the distributed photovoltaic total radiation prediction data is improved.
In order to achieve the above object, the present invention provides a method for predicting total radiation of distributed photovoltaic based on deep learning, as shown in fig. 1, including:
step 1: screening out numerical weather forecast data in a prediction time period based on a predetermined meteorological influence factor of distributed photovoltaic total radiation;
and 2, step: substituting the screened numerical weather forecast data of the prediction time interval into a pre-established total radiation prediction model based on deep learning to obtain a prediction value of the distributed photovoltaic total radiation of the prediction time interval;
the weather influence factor of the distributed photovoltaic total radiation is determined by utilizing the correlation between the historical total radiation data of the weather monitoring stations in the prediction region;
and the total radiation prediction model carries out iterative training on the numerical mode prediction data and the total radiation monitoring data by utilizing a neural network model to obtain distributed photovoltaic total radiation prediction data corresponding to a prediction time period.
The determining process of the weather influence factor of the distributed photovoltaic total radiation in the step 1 is as follows:
collecting longitude and latitude information of all distributed photovoltaic points in the prediction area, searching weather monitoring stations (including data such as total radiation, temperature, air pressure, wind speed and the like, such as a centralized photovoltaic power station weather monitoring station) closest to the area according to the longitude and latitude, collecting 1-year total radiation data, forecasting 1-year weather data corresponding to the total radiation monitoring stations by using a numerical mode, wherein each data time interval is not more than 10 minutes;
analyzing the correlation coefficient of the influence factors such as total radiation, 2m air temperature, relative humidity, sea level air pressure, 10m wind speed and the like in the meteorological data numerical weather forecast data of 1 year corresponding to the 1 year total radiation monitoring data and the numerical mode forecast total radiation monitoring station, wherein the correlation coefficient r is calculated according to the following formula:
in the formula, r x Correlation coefficient of x type meteorological element and historical total radiation monitoring data, x i Is the meteorological data of the ith time in the time series data of the x-th type meteorological elements, x is the average value of the meteorological data of each time in the time series data of the x-th type meteorological elements, y i Historical total radiation monitoring data at the ith moment, y is the average value of the historical total radiation monitoring data, and n is the total moment;
according to the correlation analysis of the total radiation meteorological influence factors (generally, the correlation coefficient is selected according to the size of the correlation coefficient, and the correlation coefficient of the total radiation meteorological influence factors is at least larger than 0.1), the total radiation, the air temperature of 2m, the relative humidity, the sea level air pressure and the wind speed of 10m are selected as the influence factors of the total radiation.
The process of constructing the total radiation prediction model based on deep learning in the step 2 is specifically as follows:
using 1 year number mode forecast meteorological element time sequence data as input, including total radiation R f 2m air temperature T f Relative humidity H f Sea level pressure P f 10m wind speed W f 1 year monitoring of total radiation data R o For output, a total radiation prediction model based on deep learning is established as shown in FIG. 2;
wherein, the deep learning of the model comprises N neurons, the excitation function of the jth hidden unit selects a Gaussian function, and the output is
j =1,2,.., N, X (t) is a set of input training samples at time t; g (t) is the center of a Gaussian function at the time t; σ is a variance of a Gaussian function; w is a j (t) j =1,2, N is the weight of the hidden layer and the output layer at the time t, and a threshold is set
The output equivalent to one hidden layer is constantly 1, which indicates that a hidden layer neuron with the output constantly 1 exists, so that a generalized network is established; in the model, the center, the variance and the weight of a basic function in the training process of deep learning are three important parameters needing learning, and the center of the basic function is determined by a self-organizing selection method based on a K-means clustering algorithm; for variance of basis function
Is determined in which d max Selecting the maximum distance between the centers of the basis functions; the weight between the hidden layer and the output layer is determined by the least square method, and the output of the deep learning is
The deep learning adopts supervised learning, and the learning of the weight coefficient of the neural network can be converted into an extremum solving problem of a multivariate linear function. In the network learning process, if the output value of the k-th iteration network is Out (k), the target value (measured value) is t p (k) Defining an objective function
Will be provided with
Adjusting the network weight coefficients according to the direction of negative gradient, i.e.
After multiple iterations, when the objective function J (k) is smaller than a certain set value, the iterations are considered to be converged, the iterations are stopped, and the weight coefficient of the network is determined, so that the training of the model is completed.
When the distributed photovoltaic total radiation needs to be predicted, firstly, according to the meteorological influence factor determined in the meteorological influence factor screening process of the distributed photovoltaic total radiation, the method in the step 1 is utilized to obtain the total radiation, the 2m air temperature, the relative humidity, the sea level air pressure and the 10m wind speed time sequence data in a prediction time period;
and (3) after data are obtained, inputting the obtained data into a total radiation prediction model based on deep learning according to the method in the step (2), and predicting a total radiation prediction result.
In order to verify the accuracy of the distributed photovoltaic total radiation prediction method based on deep learning, the method is verified by numerical weather forecast data of a Nanjing Pukou roof photovoltaic power station;
the following table 1 is a prediction error statistical table of the total radiation prediction of the Nanjing Pukou roof photovoltaic power station by using the method of the invention:
TABLE 1 Nanjing Pukou rooftop photovoltaic power station error statistics
The table shows that the method has high accuracy of the prediction effect, and is beneficial to the improvement of the prediction precision of the photovoltaic power generation power;
the average absolute error MAE index reflects an error of an absolute value of a prediction result, and can reflect an error condition to a certain extent. However, the point with a large error is easily buried when statistical averaging is performed, and the extreme case with a particularly large error cannot be reflected, and the calculation formula of the average absolute error MAE is as follows:
in the formula, Y l As a predictor of the l-th data, Z l The actual value of the ith data, and L is the number of data.
Example 2
Based on the same inventive concept, the invention provides a deep learning-based distributed photovoltaic total radiation prediction system, as shown in fig. 3, the system comprises:
the screening module is used for screening out numerical weather forecast data in a prediction time period based on a predetermined meteorological influence factor of the distributed photovoltaic total radiation;
the prediction module is used for substituting the screened numerical weather forecast data of the prediction time interval into a pre-established total radiation prediction model based on deep learning to obtain a prediction value of the distributed photovoltaic total radiation of the prediction time interval;
the weather influence factor of the distributed photovoltaic total radiation is determined by utilizing the correlation between the historical total radiation data of the weather monitoring station in the prediction region;
and the total radiation prediction model carries out iterative training on the numerical mode prediction data and the total radiation monitoring data by utilizing a neural network model to obtain distributed photovoltaic total radiation prediction data corresponding to the prediction time period.
Specifically, the process for determining the weather influence factor of the distributed photovoltaic total radiation includes:
determining a nearest weather monitoring station based on the position information of all distributed photovoltaic points in the area;
acquiring historical total radiation time sequence data of the weather station, and acquiring time sequence data of a plurality of weather elements corresponding to the historical total radiation time sequence data;
and determining correlation coefficients of the acquired time sequence data of the meteorological elements corresponding to the historical total radiation time sequence data and the historical total radiation time sequence data, and determining meteorological influence factors of the distributed photovoltaic total radiation based on the correlation coefficients.
Wherein the calculation formula of the correlation coefficient between the acquired time series data of the plurality of meteorological elements corresponding to the historical total radiation time series data and the historical total radiation time series data is as follows:
in the formula, r x Correlation coefficient of x type meteorological element and historical total radiation monitoring data, x i The meteorological data at the ith time in the time series data of the xth type meteorological element,
is the average value of meteorological data at each time in the time series data of the x-th type meteorological elements, y i Historical total radiation monitoring data for the ith time instant,
the average value of the historical total radiation monitoring data is shown, and n is the total time.
Specifically, the weather influence factor includes:
total radiation, air temperature, relative humidity, sea level air pressure and wind speed.
Specifically, the process of establishing the total radiation prediction model includes:
taking the screened numerical weather forecast data influencing the distributed photovoltaic total radiation in the historical time period as initial neural network model input data, taking total radiation monitoring data in the historical time period as initial neural network model output data, and determining the center and variance of a basis function of the initial neural network model according to the initial neural network model input data;
and performing iterative training on the weight coefficient of the initial neural network model based on the center and the variance of the basis function to obtain a deep learning total radiation prediction model.
The iterative training of the weight coefficient of the initial neural network model based on the center and the variance of the basis function to obtain the deep learning total radiation prediction model comprises the following steps:
setting an initial value of a weight coefficient of the initial neural network model;
performing iterative training on the weight coefficient of the initial neural network model based on the center, the variance and the initial value of the weight coefficient of the basis function, and determining the weight coefficient of the initial neural network during iterative convergence by using a least square method;
determining the corresponding relation between the total radiation monitoring data in the historical time period and the screened numerical weather forecast data influencing the distributed photovoltaic total radiation in the historical time period by using the weight coefficient of the initial neural network;
and determining a deep learning total radiation prediction model based on the corresponding relation between the total radiation monitoring data in the historical time period and the numerical weather forecast data which influence the distributed photovoltaic total radiation in the screened historical time period.
The center of the basis function is determined by a self-organizing selection method based on a K-means clustering algorithm.
The variance of the basis functions is determined based on the maximum distance between the centers of the basis functions.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.
Claims (10)
1. A distributed photovoltaic total radiation prediction method based on deep learning is characterized by comprising the following steps:
screening out numerical weather forecast data in a prediction time period based on a predetermined meteorological influence factor of distributed photovoltaic total radiation;
substituting the screened numerical weather forecast data of the prediction time interval into a pre-established total radiation prediction model based on deep learning to obtain a prediction value of the distributed photovoltaic total radiation of the prediction time interval;
the meteorological influence factor of the distributed photovoltaic total radiation is determined by utilizing the correlation between the historical total radiation data of the meteorological monitoring station in the forecast area;
and the total radiation prediction model carries out iterative training on the numerical mode prediction data and the total radiation monitoring data by utilizing a neural network model to obtain distributed photovoltaic total radiation prediction data corresponding to a prediction time period.
2. The method of claim 1, wherein the determining of the weather-affecting factor for the distributed photovoltaic total radiation comprises:
determining a nearest weather monitoring station based on the position information of all distributed photovoltaic points in the area;
acquiring historical total radiation time sequence data of the weather station, and acquiring time sequence data of a plurality of weather elements corresponding to the historical total radiation time sequence data;
and determining correlation coefficients of the acquired time sequence data of the meteorological elements corresponding to the historical total radiation time sequence data and the historical total radiation time sequence data, and determining meteorological influence factors of the distributed photovoltaic total radiation based on the correlation coefficients.
3. The method according to claim 2, wherein the correlation coefficient between the acquired time series data of the plurality of meteorological elements corresponding to the historical total radiation time series data and the historical total radiation time series data is calculated as follows:
in the formula, r x Correlation coefficient of x type meteorological element and historical total radiation monitoring data, x i The meteorological data at the ith moment in the time series data of the meteorological elements of the x type,
is the average value of meteorological data at each time in the time series data of the x-th type meteorological elements, y i Historical total radiation monitoring data for the ith time instant,
the average value of the historical total radiation monitoring data is shown, and n is the total time.
4. The method of any of claims 1 to 2, wherein the weather affecting factor comprises:
total radiation, air temperature, relative humidity, sea level air pressure and wind speed.
5. The method of claim 1, wherein the building of the total radiance prediction model comprises:
taking the screened numerical weather forecast data influencing the distributed photovoltaic total radiation in the historical time period as initial neural network model input data, taking total radiation monitoring data in the historical time period as initial neural network model output data, and determining the center and variance of a basis function of the initial neural network model according to the initial neural network model input data;
and performing iterative training on the weight coefficient of the initial neural network model based on the center and the variance of the basis function to obtain a deep learning total radiation prediction model.
6. The method of claim 5, wherein iteratively training the weight coefficients of the initial neural network model based on the centers and variances of the basis functions to obtain a deep-learning total radiation prediction model comprises:
setting an initial value of a weight coefficient of the initial neural network model;
performing iterative training on the weight coefficient of the initial neural network model based on the center, the variance and the initial value of the weight coefficient of the basis function, and determining the weight coefficient of the initial neural network during iterative convergence by using a least square method;
determining the corresponding relation between the total radiation monitoring data in the historical time period and the screened numerical weather forecast data influencing the distributed photovoltaic total radiation in the historical time period by using the weight coefficient of the initial neural network;
and determining a deep learning total radiation prediction model based on the corresponding relation between the total radiation monitoring data in the historical time period and the numerical weather forecast data which influences the distributed photovoltaic total radiation in the screened historical time period.
7. The method of claim 5, wherein the centers of the basis functions are determined based on a self-organizing selection method of a K-means clustering algorithm.
8. The method of claim 5, wherein the variance of the basis functions is determined based on a maximum distance between the centers of the basis functions.
9. A deep learning based distributed photovoltaic total radiation prediction system, the system comprising:
the screening module is used for screening out numerical weather forecast data in a prediction time period based on a predetermined meteorological influence factor of the distributed photovoltaic total radiation;
the prediction module is used for substituting the screened numerical weather forecast data of the prediction time interval into a pre-established total radiation prediction model based on deep learning to obtain a prediction value of the distributed photovoltaic total radiation of the prediction time interval;
the weather influence factor of the distributed photovoltaic total radiation is determined by utilizing the correlation between the historical total radiation data of the weather monitoring stations in the prediction region;
and the total radiation prediction model carries out iterative training on the numerical mode prediction data and the total radiation monitoring data by utilizing a neural network model to obtain distributed photovoltaic total radiation prediction data corresponding to a prediction time period.
10. The system of claim 9, wherein the process of determining the weather-affecting factor for the distributed photovoltaic total radiation comprises:
determining a nearest weather monitoring station based on the position information of all distributed photovoltaic points in the area;
acquiring historical total radiation time sequence data of the weather station, and acquiring time sequence data of a plurality of weather elements corresponding to the historical total radiation time sequence data;
and determining correlation coefficients of the acquired time sequence data of the meteorological elements corresponding to the historical total radiation time sequence data and the historical total radiation time sequence data, and determining meteorological influence factors of the distributed photovoltaic total radiation based on the correlation coefficients.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110741009.8A CN115563848A (en) | 2021-07-01 | 2021-07-01 | Distributed photovoltaic total radiation prediction method and system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110741009.8A CN115563848A (en) | 2021-07-01 | 2021-07-01 | Distributed photovoltaic total radiation prediction method and system based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115563848A true CN115563848A (en) | 2023-01-03 |
Family
ID=84737820
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110741009.8A Pending CN115563848A (en) | 2021-07-01 | 2021-07-01 | Distributed photovoltaic total radiation prediction method and system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115563848A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116760031A (en) * | 2023-08-17 | 2023-09-15 | 北京弘象科技有限公司 | High-time-resolution photovoltaic power prediction method and device based on meteorological data |
-
2021
- 2021-07-01 CN CN202110741009.8A patent/CN115563848A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116760031A (en) * | 2023-08-17 | 2023-09-15 | 北京弘象科技有限公司 | High-time-resolution photovoltaic power prediction method and device based on meteorological data |
| CN116760031B (en) * | 2023-08-17 | 2023-10-27 | 北京弘象科技有限公司 | High-time-resolution photovoltaic power prediction method and device based on meteorological data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103268366A (en) | Combined wind power prediction method suitable for distributed wind power plant | |
| CN109143408B (en) | Dynamic region combined short-time rainfall forecasting method based on MLP | |
| CN111833202B (en) | Farmland evapotranspiration short-term prediction method considering crop coefficient dynamic change and rainfall | |
| CN112036658B (en) | Wind power real-time prediction method based on forgetting long-term and short-term memory with difference | |
| de Guia et al. | Using stacked long short term memory with principal component analysis for short term prediction of solar irradiance based on weather patterns | |
| CN116595394A (en) | Training method of wind speed correction model, wind speed prediction method, wind speed prediction equipment and medium | |
| CN112288157A (en) | Wind power plant power prediction method based on fuzzy clustering and deep reinforcement learning | |
| CN117117819A (en) | Photovoltaic power generation short-term power prediction method, system, equipment and medium | |
| CN114357670A (en) | Power distribution network power consumption data abnormity early warning method based on BLS and self-encoder | |
| CN115563848A (en) | Distributed photovoltaic total radiation prediction method and system based on deep learning | |
| CN117526274A (en) | New energy power prediction method, electronic equipment and storage medium in extreme climate | |
| CN112307672A (en) | BP neural network short-term wind power prediction method based on cuckoo algorithm optimization | |
| CN113723670B (en) | Photovoltaic power generation power short-term prediction method with variable time window | |
| CN116484998A (en) | Distributed photovoltaic power station power prediction method and system based on meteorological similar day | |
| CN114912716A (en) | Short-term power load prediction method and system based on bidirectional long-term and short-term memory network | |
| CN115907131A (en) | Method and system for building electric heating load prediction model in northern area | |
| CN114897204A (en) | Method and device for predicting short-term wind speed of offshore wind farm | |
| CN114862023A (en) | Distributed photovoltaic power prediction method and system based on four-dimensional point-by-point meteorological forecast | |
| CN114676866A (en) | Method, device and storage medium for correcting load prediction before day based on error correction | |
| CN113780644A (en) | Photovoltaic output prediction method based on online learning | |
| CN113159395A (en) | Deep learning-based sewage treatment plant water inflow prediction method and system | |
| CN112215383A (en) | Distributed photovoltaic power generation power prediction method and system | |
| CN113887843B (en) | Group prediction method and system for wind power plant output power | |
| CN116960960B (en) | Short-term wind power prediction method for wind turbine | |
| CN116822658B (en) | Early warning method and device for wind-solar resources, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication |