CN112541633A - Ultrahigh-precision photovoltaic power generation power prediction method and system based on neural network - Google Patents
Ultrahigh-precision photovoltaic power generation power prediction method and system based on neural network Download PDFInfo
- Publication number
- CN112541633A CN112541633A CN202011482191.1A CN202011482191A CN112541633A CN 112541633 A CN112541633 A CN 112541633A CN 202011482191 A CN202011482191 A CN 202011482191A CN 112541633 A CN112541633 A CN 112541633A
- Authority
- CN
- China
- Prior art keywords
- layer
- output
- neural network
- data
- predicted
- 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
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 29
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000010248 power generation Methods 0.000 title claims abstract description 23
- 238000003062 neural network model Methods 0.000 claims abstract description 35
- 238000012549 training Methods 0.000 claims abstract description 11
- 230000003287 optical effect Effects 0.000 claims description 7
- 230000015572 biosynthetic process Effects 0.000 claims description 6
- 238000003786 synthesis reaction Methods 0.000 claims description 6
- 238000012546 transfer Methods 0.000 claims description 6
- 238000012935 Averaging Methods 0.000 claims description 3
- 230000008901 benefit Effects 0.000 abstract description 5
- 238000004364 calculation method Methods 0.000 description 8
- 238000007792 addition Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 5
- 238000012217 deletion Methods 0.000 description 4
- 230000037430 deletion Effects 0.000 description 4
- 230000005284 excitation Effects 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000005684 electric field Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000007619 statistical method Methods 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
- 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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- General Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Educational Administration (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an ultrahigh-precision photovoltaic power generation power prediction method and system based on a neural network, which comprises the following steps of: establishing an EMD-Elman neural network model; collecting a plurality of groups of data associated with days to be predicted in a historical data set, establishing a similar sequence group, and training a neural network model; inputting the characteristics of the days to be predicted into the trained neural network model to obtain the predicted values of all components, and summing the predicted values to obtain the prediction result of the current similar day. The photovoltaic power prediction system adopts a physical model and a unique mathematical modeling mode, can automatically calculate related parameters, quickly and effectively adjusts a prediction link according to the actual situation on site, and provides a reliable basis for economic benefits and operation management of an owner. Short-term power prediction and ultra-short-term power prediction of the photovoltaic power station are completed by collecting numerical weather forecast data, real-time weather station data, real-time output power data and inverter unit state data, and the short-term power prediction and ultra-short-term power prediction work is completed by uploading the data to a scheduling side power prediction system according to the requirements of a power grid.
Description
Technical Field
The invention relates to the field of power generation power prediction, in particular to a method and a system for predicting ultrahigh-precision photovoltaic power generation power based on a neural network.
Background
With the development of the internet of things and the wide application of intelligent equipment, the quality of optical power data is greatly improved, and the data is more accurate and reliable than the previous manual data copying method, but the problems of data loss or data abnormity and the like still exist. Such as data loss and data anomalies due to equipment failure. The data imperfection of the database caused by these problems affects the prediction effect of the required optical power, so in order to reduce the prediction error, the missing data must be filled, and the abnormal data must be identified and corrected to obtain a set of more complete data meeting the input requirement of the prediction algorithm. Therefore, the accuracy of the prediction is improved by ensuring the completeness and accuracy of the database data.
In the prior art, a BP neural network is generally used in photovoltaic power generation power prediction, a dynamic problem is substantially converted into a static problem to be researched, and the final prediction precision is influenced.
Based on the situation, the invention provides an ultrahigh-precision photovoltaic power generation power prediction method and system based on a neural network, and the problems can be effectively solved.
Disclosure of Invention
The invention aims to provide an ultrahigh-precision photovoltaic power generation power prediction method and system based on a neural network. According to the ultrahigh-precision photovoltaic power generation power prediction method and system based on the neural network, the EMD-Elman neural network model is used as a calculation framework, so that the method and system have good stability and strong calculation capacity; and the network model is trained according to historical data, so that better accuracy of prediction is ensured.
The invention is realized by the following technical scheme:
an ultrahigh-precision photovoltaic power generation power prediction method based on a neural network comprises the following steps:
step S1: establishing an EMD-Elman neural network model;
step S2: collecting a plurality of groups of data associated with days to be predicted in a historical data set, establishing a similar sequence group, and training a neural network model;
step S3: inputting the characteristics of the days to be predicted into the trained neural network model to obtain the predicted values of all components, and summing the predicted values to obtain the prediction result of the current similar day.
The invention adopts the EMD-Elman neural network model as a calculation framework, and has better stability and stronger calculation capability; and the network model is trained according to historical data, so that better accuracy of prediction is ensured.
Preferably, the step S1 specifically includes the following steps:
step S11: respectively establishing an input layer, a hidden layer and an output layer, and sleeving a carrying layer outside the input layer, the hidden layer and the output layer;
step S12: initializing weights of an input layer, a hidden layer, an output layer and a receiving layer, inputting sample data to the input layer, and transmitting the data to the hidden layer;
step S13: calculating the output quantity of the bearing layer at the k moment according to the output quantity of the hidden layer at the k-1 moment, wherein the output quantity of the output layer is determined according to the following formula (1):
xc(k)＝x(k-1) (1)
in the formula (1), x (k-1) represents the output vector of the hidden layer at the moment of k-1, and xc(k) An output vector expression representing a lower receiving layer at the time k;
step S14: the output x of the receiving layer at time k obtained in step S13c(k) MeterCalculating the output quantity of the hidden layer at the time k, and transmitting the output quantity x (k) of the hidden layer obtained at the time k to the output layer and the receiving layer, wherein the output quantity of the hidden layer is determined according to the following formula (2):
x(k)＝f[w31xc(k)+w12(u(k-1))] (2)
in the formula (2), x (k) represents an output vector expression of the hidden layer at the time k, w _31 represents a connection weight from the receiving layer to the input layer, and w12Represents the connection weight from the input layer to the intermediate layer, u (k-1) represents the input vector, xc(k) An output vector expression of the lower receiving layer at the moment k is represented, and f (×) represents a nonlinear vector synthesis function of the intermediate layer transfer function combination;
step S15: calculating the output quantity of the output layer at the time k according to the output quantity x (k) of the hidden layer at the time k in the step S14, wherein the output quantity of the output layer is determined according to the following formula (3):
y(k)＝g(w23x(k)] (3)
in the above formula (3), w23Representing the connection weight from the middle layer to the output layer, k representing the current time, x (k) representing the output vector expression of the hidden layer at the time of k, g [. X ]]A non-linear vector synthesis function representing a combination of output layer transfer functions.
The network model has the capability of adapting to time-varying characteristics, enhances the global stability of the network model, has stronger computing capability than a feedforward neural network, and can be used for solving the problem of quick optimization.
The hidden layer unit has two types of linear and nonlinear excitation functions, and usually the excitation function is a sigmoid nonlinear function. The receiving layer is used to memorize the output value of the hidden layer unit at the previous moment, and can be regarded as a delay operator with a step delay. The output of the hidden layer is self-connected to the input of the hidden layer through the delay and storage of the bearing layer, the self-connection mode enables the hidden layer to have sensitivity to historical data, and the addition of the internal feedback network increases the capability of the network for processing dynamic information, so that the purpose of dynamic modeling is achieved.
Preferably, the step S2 specifically includes the following steps:
step S21: searching N groups of similar day sequences in the historical data set, wherein the sequences are respectively sequence 1: (U)1，Y1) And sequence 2: (U)2，Y2) And sequence 3: (U)3，Y3) … sequence N: (U)N，YN))；
Step S22: defining equation (4) as an error function of the EMD-Elman neural network:
in the above formula (4), yk(i) The target value of the output vector under the i time node is represented, and y (i) the actual value of the output vector under the i time node is represented;
step S23: selecting any group of sequences N (U)N，YN) (ii) a Will input vector UNInputting the data to an input layer of an EMD-Elman neural network for target training, and calculating to obtain an output vector ONAgain, a global error function E is obtained2See formula (5):
step S24: according to E in step S221And E in step S232To obtain the final error function, see formula (6):
preferably, the step S3 specifically includes the following steps:
step S31: searching N groups of date data similar to the days to be predicted in the historical data set, and establishing an EMD-Elman neural network model for each group of similar data;
step S32, inputting meteorological features as a network model, outputting the meteorological features as optical power prediction data, and obtaining N groups of EMD-Elman neural network models;
step S33, respectively taking the meteorological features of the days to be predicted as the input of the five groups of network models to obtain the five groups of predicted output power of the days to be predicted; and finally, summing and averaging the five groups of predicted values to obtain the final predicted output power.
The invention also provides a system for realizing the ultrahigh-precision photovoltaic power generation power prediction method based on the neural network, which comprises the following modules:
a first module for building an EMD-Elman neural network model;
the second module is used for collecting a plurality of groups of data related to the days to be predicted in the historical data set, establishing a similar sequence group and training a neural network model;
and the third module is used for inputting the characteristics of the day to be predicted into the trained neural network model to obtain the predicted values of all the components, and summing the predicted values to obtain the prediction result of the current similar day.
Preferably, the first module includes an input layer, a hidden layer, an output layer, and a socket layer.
Compared with the prior art, the invention has the following advantages and beneficial effects:
according to the ultrahigh-precision photovoltaic power generation power prediction method and system based on the neural network, the EMD-Elman neural network model is used as a calculation framework, so that the method and system have good stability and strong calculation capacity; and the network model is trained according to historical data, so that better accuracy of prediction is ensured.
The photovoltaic power prediction system adopts a physical model and a unique mathematical modeling mode, can automatically calculate related parameters, can quickly and effectively adjust a prediction link according to the actual situation on site, reduces light abandonment, and provides a reliable basis for economic benefits and operation management of owners. Short-term power prediction and ultra-short-term power prediction of the photovoltaic power station are completed by collecting data such as numerical weather forecast data, real-time weather station data, real-time output power data and inverter unit state, and the short-term power prediction and ultra-short-term power prediction work is completed by uploading the data to a scheduling side power prediction system according to the requirements of a power grid.
Drawings
FIG. 1 is a flow chart of the operation of the present invention;
FIG. 2 is a schematic diagram of the EMD-Elman neural network model of the present invention;
FIG. 3 is a flow chart of training a neural network model according to the present invention.
Detailed Description
In order that those skilled in the art will better understand the technical solutions of the present invention, the following description of the preferred embodiments of the present invention is provided in conjunction with specific examples, but it should be understood that the drawings are for illustrative purposes only and should not be construed as limiting the patent; for the purpose of better illustrating the embodiments, certain features of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product; it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted. The positional relationships depicted in the drawings are for illustrative purposes only and are not to be construed as limiting the present patent.
Example 1:
as shown in fig. 1 to 3, an ultra-high precision photovoltaic power generation power prediction method based on a neural network includes the following steps:
step S1: establishing an EMD-Elman neural network model;
step S2: collecting a plurality of groups of data associated with days to be predicted in a historical data set, establishing a similar sequence group, and training a neural network model;
step S3: inputting the characteristics of the days to be predicted into the trained neural network model to obtain the predicted values of all components, and summing the predicted values to obtain the prediction result of the current similar day.
The invention adopts the EMD-Elman neural network model as a calculation framework, and has better stability and stronger calculation capability; and the network model is trained according to historical data, so that better accuracy of prediction is ensured.
Further, in another embodiment, the step S1 specifically includes the following steps:
step S11: respectively establishing an input layer, a hidden layer and an output layer, and sleeving a carrying layer outside the input layer, the hidden layer and the output layer;
step S12: initializing weights of an input layer, a hidden layer, an output layer and a receiving layer, inputting sample data to the input layer, and transmitting the data to the hidden layer;
step S13: calculating the output quantity of the bearing layer at the k moment according to the output quantity of the hidden layer at the k-1 moment, wherein the output quantity of the output layer is determined according to the following formula (1):
xc(k)＝x(k-1) (1)
in the formula (1), x (k-1) represents the output vector of the hidden layer at the moment of k-1, and xc(k) An output vector expression representing a lower receiving layer at the time k;
step S14: the output x of the receiving layer at time k obtained in step S13c(k) Calculating the output quantity of the hidden layer at the time k, and transmitting the hidden layer output quantity x (k) obtained at the time k to the output layer and the receiving layer, wherein the output quantity of the hidden layer is determined according to the following formula (2):
x(k)＝f[w31xc(k)+w12(u(k-1))] (2)
in the formula (2), x (k) represents an output vector expression of the hidden layer at the time k, w _31 represents a connection weight from the receiving layer to the input layer, and w12Represents the connection weight from the input layer to the intermediate layer, u (k-1) represents the input vector, xc(k) An output vector expression of the lower receiving layer at the moment k is represented, and f (×) represents a nonlinear vector synthesis function of the intermediate layer transfer function combination;
step S15: calculating the output quantity of the output layer at the time k according to the output quantity x (k) of the hidden layer at the time k in the step S14, wherein the output quantity of the output layer is determined according to the following formula (3):
y(k)＝g[w23x(k)] (3)
in the above formula (3), w23Representing the connection weight from the middle layer to the output layer, k representing the current time, x (k) representing the output vector expression of the hidden layer at the time of k, g [. X ]]A non-linear vector synthesis function representing a combination of output layer transfer functions.
The network model has the capability of adapting to time-varying characteristics, enhances the global stability of the network model, has stronger computing capability than a feedforward neural network, and can be used for solving the problem of quick optimization.
The hidden layer unit has two types of linear and nonlinear excitation functions, and usually the excitation function is a sigmoid nonlinear function. The receiving layer is used to memorize the output value of the hidden layer unit at the previous moment, and can be regarded as a delay operator with a step delay. The output of the hidden layer is self-connected to the input of the hidden layer through the delay and storage of the bearing layer, the self-connection mode enables the hidden layer to have sensitivity to historical data, and the addition of the internal feedback network increases the capability of the network for processing dynamic information, so that the purpose of dynamic modeling is achieved.
Further, in another embodiment, the step S2 specifically includes the following steps:
step S21: searching N groups of similar day sequences in the historical data set, wherein the sequences are respectively sequence 1: (U)1，Y1) And sequence 2: (U)2，Y2) And sequence 3: (U)3，Y3) … sequence N: (U)N，YN)；
Step S22: defining equation (4) as an error function of the EMD-Elman neural network:
in the above formula (4), yk(i) The target value of the output vector under the i time node is represented, and y (i) the actual value of the output vector under the i time node is represented;
step S23: selecting any group of sequences N (U)N，YN) (ii) a Will input vector UNInputting the data to an input layer of an EMD-Elman neural network for target training, and calculating to obtain an output vector ONAgain, a global error function E is obtained2See formula (5):
step S24: according to E in step S221And E in step S232To obtain the final error function, see formula (6):
further, in another embodiment, the step S3 specifically includes the following steps:
step S31: searching N groups of date data similar to the days to be predicted in the historical data set, and establishing an EMD-Elman neural network model for each group of similar data;
step S32, inputting meteorological features as a network model, outputting the meteorological features as optical power prediction data, and obtaining N groups of EMD-Elman neural network models;
step S33, respectively taking the meteorological features of the days to be predicted as the input of the five groups of network models to obtain the five groups of predicted output power of the days to be predicted; and finally, summing and averaging the five groups of predicted values to obtain the final predicted output power.
When the generated power is predicted, the mathematical model method has the advantages of small calculated amount and high speed, but has a lot of defects and limitations, such as lack of self-learning and self-adaptive capacity, no guarantee on robustness of a prediction system, and the like. Particularly, with the development of economy in China, the structure of a power system is increasingly complex, the characteristics of nonlinearity, time-varying property and uncertainty of power load change are more obvious, and a proper mathematical model is difficult to establish to clearly express the relationship between the load and the variable influencing the load. This problem is exacerbated as power systems become increasingly complex.
With the development of the internet of things and the wide application of intelligent equipment, the quality of optical power data is greatly improved, and the data is more accurate and reliable than the previous manual data copying method, but the problems of data loss or data abnormity and the like still exist. Such as data loss and data anomalies due to equipment failure. The data imperfection of the database caused by these problems affects the prediction effect of the required optical power, so in order to reduce the prediction error, the missing data must be filled, and the abnormal data must be identified and corrected to obtain a set of more complete data meeting the input requirement of the prediction algorithm. Therefore, the accuracy of the prediction is improved by ensuring the completeness and accuracy of the database data.
The photovoltaic power prediction system adopts a physical model and a unique mathematical modeling mode, can automatically calculate related parameters, can quickly and effectively adjust a prediction link according to the actual situation on site, reduces light abandonment, and provides a reliable basis for economic benefits and operation management of owners. The EMD-Elman neural network combined prediction model first looks for five groups of similar day sequences in the historical dataset through grey correlation analysis. The sequence 1 is that the similar day 1 is most similar to the day to be predicted, and the relevance of other similar days is decreased in sequence. Then EMD decomposition is carried out on the similar sunlight power to obtain an eigenmode function and a residual component. The similar solar meteorological features are used as input of the Elman neural network model, the output is an eigenmode function and a residual component, and the components are corresponding to the Elman neural network model. And finally, inputting the solar meteorological features to be predicted as a trained model to obtain predicted values of all components, and summing the predicted values to obtain a prediction result of a similar day 1. And in the same way, five groups of prediction results are summed and averaged to obtain the final prediction result.
The invention further provides a system for realizing the ultrahigh-precision photovoltaic power generation power prediction method based on the neural network, which comprises the following modules:
a first module for building an EMD-Elman neural network model;
the second module is used for collecting a plurality of groups of data related to the days to be predicted in the historical data set, establishing a similar sequence group and training a neural network model;
and the third module is used for inputting the characteristics of the day to be predicted into the trained neural network model to obtain the predicted values of all the components, and summing the predicted values to obtain the prediction result of the current similar day.
Further, in another embodiment, the first module includes an input layer, an hidden layer, an output layer, and a socket layer.
The prediction system of the invention has the functions of data processing, power prediction, interface query statistics and the like.
Data acquisition and processing:
the data required by the operation of the photovoltaic power generation power prediction system comprise numerical weather forecast data, automatic environment monitoring station data, photovoltaic power station real-time output power data, photovoltaic module operation state and the like.
The system can complete automatic acquisition of all data and can also manually input the data, wherein data forecast data is automatically acquired at regular time. The acquisition cycle of the real-time power data of the photovoltaic power station is not more than 1min, and the acquisition cycle of the state data of the photovoltaic module and the inverter is not more than 15 min.
The data statistics function is as follows:
a. the time range of the system participating in the statistical data can be selected at will;
b. historical power data statistics, historical meteorological data statistics, and the like;
c. the system counts errors between prediction results and actual power generation amount of various parameters under different conditions, and the statistical results can be dynamically generated and exported for use in subsequent analysis, scientific research and other work;
d. the system can perform error analysis on the prediction result in any time interval, and the error indexes comprise mean square error, mean absolute error, correlation coefficient and the like.
Data analysis and processing functions:
a. the prediction system has a series of screening and processing functions on basic data, and reduces the interference of abnormal values and error information on prediction results;
b. the forecasting system can compare and analyze the weather forecast value and the data of the environment monitoring station;
c. parameter interaction is realized among prediction systems, and the correlation degree among different parameters is analyzed.
And (3) data output:
according to the power grid dispatching requirement, short-term and ultra-short-term power prediction in a standard format, real-time monitoring of an automatic environment monitoring station, inverter overhaul capacity, installed capacity of a photovoltaic power station, maximum output and other information reporting can be achieved. Real-time radiation degree, ambient temperature, photovoltaic module temperature and other meteorological information of the photovoltaic power station need to be uploaded to a power grid dispatching mechanism through teleoperation.
And (3) starting a system:
the short-term photovoltaic power prediction can be carried out once every 15min, and automatic rolling prediction is realized
A system graphical interface:
the system graphical interface is divided into 8 modules according to functions: the system comprises a user management module, a system setting module, a state monitoring module, a prediction curve module, a meteorological information module, a statistical analysis module, a data report module and a help module.
The user management module has the functions of adding users, modifying user information, deleting users and the like.
The system setting comprises the addition, modification and deletion of electric field basic information, the addition, modification and deletion of inverter information, the addition, modification and deletion of a weather station, the addition, modification and deletion of a photovoltaic module, the setting of prediction information and the power limiting setting.
According to the description and the drawings of the invention, the ultrahigh-precision photovoltaic power generation power prediction method and system based on the neural network can be easily manufactured or used by a person skilled in the art, and can generate the positive effects recorded in the invention.
Unless otherwise specified, in the present invention, if there is an orientation or positional relationship indicated by terms of "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., based on the orientation or positional relationship shown in the drawings, it is only for convenience of describing the present invention and simplifying the description, rather than to indicate or imply that the device or element so referred to must have a particular orientation, be constructed and operated in a particular orientation, therefore, the terms describing orientation or positional relationship in the present invention are for illustrative purposes only, and should not be construed as limiting the present patent, specific meanings of the above terms can be understood by those of ordinary skill in the art in light of the specific circumstances in conjunction with the accompanying drawings.
Unless expressly stated or limited otherwise, the terms "disposed," "connected," and "connected" are used broadly and encompass, for example, being fixedly connected, detachably connected, or integrally connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention in any way, and all simple modifications and equivalent variations of the above embodiments according to the technical spirit of the present invention are included in the scope of the present invention.
Claims (6)
1. A method for predicting ultrahigh-precision photovoltaic power generation power based on a neural network is characterized by comprising the following steps: the method comprises the following steps:
step S1: establishing an EMD-Elman neural network model;
step S2: collecting a plurality of groups of data associated with days to be predicted in a historical data set, establishing a similar sequence group, and training a neural network model;
step S3: inputting the characteristics of the days to be predicted into the trained neural network model to obtain the predicted values of all components, and summing the predicted values to obtain the prediction result of the current similar day.
2. The ultrahigh-precision photovoltaic power generation power prediction method based on the neural network as claimed in claim 1, wherein: the step S1 specifically includes the following steps:
step S11: respectively establishing an input layer, a hidden layer and an output layer, and sleeving a carrying layer outside the input layer, the hidden layer and the output layer;
step S12: initializing weights of an input layer, a hidden layer, an output layer and a receiving layer, inputting sample data to the input layer, and transmitting the data to the hidden layer;
step S13: calculating the output quantity of the bearing layer at the k moment according to the output quantity of the hidden layer at the k-1 moment, wherein the output quantity of the output layer is determined according to the following formula (1):
xc(k)＝x(k-1) (1)
in the formula (1), x (k-1) represents the output vector of the hidden layer at the moment of k-1, and xc(k) An output vector expression representing a lower receiving layer at the time k;
step S14: the output x of the receiving layer at time k obtained in step S13c(k) Calculating the output quantity of the hidden layer at the time k, and transmitting the hidden layer output quantity x (k) obtained at the time k to the output layer and the receiving layer, wherein the output quantity of the hidden layer is determined according to the following formula (2):
x(k)＝f[w31xc(k)+w12(u(k-1))] (2)
in the formula (2), x (k) represents an output vector expression of the hidden layer at the time k, w _31 represents a connection weight from the receiving layer to the input layer, and w12Represents the connection weight from the input layer to the intermediate layer, u (k-1) represents the input vector, xc(k) An output vector expression of the lower receiving layer at the moment k is represented, and f (×) represents a nonlinear vector synthesis function of the intermediate layer transfer function combination;
step S15: calculating the output quantity of the output layer at the time k according to the output quantity x (k) of the hidden layer at the time k in the step S14, wherein the output quantity of the output layer is determined according to the following formula (3):
y(k)＝g[w23x(k)] (3)
in the above formula (3), w23Representing the connection weight from the middle layer to the output layer, k representing the current time, x (k) representing the output vector expression of the hidden layer at the time of k, g [. X ]]A non-linear vector synthesis function representing a combination of output layer transfer functions.
3. The ultrahigh-precision photovoltaic power generation power prediction method based on the neural network as claimed in claim 2, wherein: the step S2 specifically includes the following steps:
step S21: searching N groups of similar day sequences in the historical data set, wherein the sequences are respectively sequence 1: (U)1，Y1) Sequence of2：(U2，Y2) And sequence 3: (U)3，Y3) … sequence N: (U)N，YN)；
Step S22: defining equation (4) as an error function of the EMD-Elman neural network:
in the above formula (4), yk(i) The target value of the output vector under the i time node is represented, and y (i) the actual value of the output vector under the i time node is represented;
step S23: selecting any group of sequences N (U)N，YN) (ii) a Will input vector UNInputting the data to an input layer of an EMD-Elman neural network for target training, and calculating to obtain an output vector ONAgain, a global error function E is obtained2See formula (5):
step S24: according to E in step S221And E in step S232To obtain the final error function, see formula (6):
4. the ultrahigh-precision photovoltaic power generation power prediction method based on the neural network as claimed in claim 3, wherein: the step S3 specifically includes the following steps:
step S31: searching N groups of date data similar to the days to be predicted in the historical data set, and establishing an EMD-Elman neural network model for each group of similar data;
step S32, inputting meteorological features as a network model, outputting the meteorological features as optical power prediction data, and obtaining N groups of EMD-Elman neural network models;
step S33, respectively taking the meteorological features of the days to be predicted as the input of the five groups of network models to obtain the five groups of predicted output power of the days to be predicted; and finally, summing and averaging the five groups of predicted values to obtain the final predicted output power.
5. A system for realizing the ultrahigh-precision photovoltaic power generation power prediction method based on the neural network according to any one of claims 1 to 4, wherein the method comprises the following steps: the system comprises the following modules:
a first module for building an EMD-Elman neural network model;
the second module is used for collecting a plurality of groups of data related to the days to be predicted in the historical data set, establishing a similar sequence group and training a neural network model;
and the third module is used for inputting the characteristics of the day to be predicted into the trained neural network model to obtain the predicted values of all the components, and summing the predicted values to obtain the prediction result of the current similar day.
6. The ultra-high precision photovoltaic power generation power prediction system based on the neural network as claimed in claim 5, wherein: the first module includes an input layer, a hidden layer, an output layer, and a receiving layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011482191.1A CN112541633A (en) | 2020-12-15 | 2020-12-15 | Ultrahigh-precision photovoltaic power generation power prediction method and system based on neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011482191.1A CN112541633A (en) | 2020-12-15 | 2020-12-15 | Ultrahigh-precision photovoltaic power generation power prediction method and system based on neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112541633A true CN112541633A (en) | 2021-03-23 |
Family
ID=75018798
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011482191.1A Pending CN112541633A (en) | 2020-12-15 | 2020-12-15 | Ultrahigh-precision photovoltaic power generation power prediction method and system based on neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112541633A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114123200A (en) * | 2022-01-24 | 2022-03-01 | 国网山西省电力公司晋城供电公司 | Photovoltaic power station dynamic modeling method based on data driving and storage device |
| CN116011686A (en) * | 2023-03-28 | 2023-04-25 | 山东创宇环保科技有限公司 | Charging shed photovoltaic power generation reserve prediction method based on multi-data fusion |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104268638A (en) * | 2014-09-11 | 2015-01-07 | 广州市香港科大霍英东研究院 | Photovoltaic power generation system power predicting method of elman-based neural network |
| CN106295798A (en) * | 2016-08-29 | 2017-01-04 | 江苏省电力试验研究院有限公司 | Empirical mode decomposition and Elman neural network ensemble wind-powered electricity generation Forecasting Methodology |
| CN109978283A (en) * | 2019-04-25 | 2019-07-05 | 中国人民解放军国防科技大学 | Photovoltaic power generation power prediction method based on branch evolution neural network |
-
2020
- 2020-12-15 CN CN202011482191.1A patent/CN112541633A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104268638A (en) * | 2014-09-11 | 2015-01-07 | 广州市香港科大霍英东研究院 | Photovoltaic power generation system power predicting method of elman-based neural network |
| CN106295798A (en) * | 2016-08-29 | 2017-01-04 | 江苏省电力试验研究院有限公司 | Empirical mode decomposition and Elman neural network ensemble wind-powered electricity generation Forecasting Methodology |
| CN109978283A (en) * | 2019-04-25 | 2019-07-05 | 中国人民解放军国防科技大学 | Photovoltaic power generation power prediction method based on branch evolution neural network |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114123200A (en) * | 2022-01-24 | 2022-03-01 | 国网山西省电力公司晋城供电公司 | Photovoltaic power station dynamic modeling method based on data driving and storage device |
| CN116011686A (en) * | 2023-03-28 | 2023-04-25 | 山东创宇环保科技有限公司 | Charging shed photovoltaic power generation reserve prediction method based on multi-data fusion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhao et al. | A novel bidirectional mechanism based on time series model for wind power forecasting | |
| CN113128793A (en) | Photovoltaic power combination prediction method and system based on multi-source data fusion | |
| CN112734128B (en) | 7-day power load peak prediction method based on optimized RBF | |
| CN112712189B (en) | Heat supply demand load prediction method | |
| CN116150897A (en) | Machine tool spindle performance evaluation method and system based on digital twin | |
| CN112541633A (en) | Ultrahigh-precision photovoltaic power generation power prediction method and system based on neural network | |
| Jiang et al. | Day‐ahead renewable scenario forecasts based on generative adversarial networks | |
| Boumaiza et al. | Modeling multi-criteria decision analysis in residential PV adoption | |
| CN111340305A (en) | Building operation energy consumption prediction method | |
| CN107145968A (en) | Photovoltaic apparatus life cycle cost Forecasting Methodology and system based on BP neural network | |
| CN116523187A (en) | Engineering progress monitoring method and system based on BIM | |
| CN116029419A (en) | Deep learning-based long-term new energy daily average generation power prediction method and system | |
| CN114429248A (en) | Transformer apparent power prediction method | |
| CN111222762A (en) | Solar cell panel coating process state monitoring and quality control system | |
| CN108108837B (en) | Regional new energy power supply structure optimization prediction method and system | |
| Ding et al. | A big data based cost prediction method for remanufacturing end-of-life products | |
| Tian | Adaptive control and supply chain management of intelligent agricultural greenhouse by intelligent fuzzy auxiliary cognitive system | |
| CN112801388B (en) | Power load prediction method and system based on nonlinear time series algorithm | |
| CN111461297B (en) | Solar irradiation quantity optimization prediction algorithm based on MPC and ELM neural network | |
| Kopyltsov et al. | Algorithm of estimation and correction of wireless telecommunications quality | |
| Kariniotakis et al. | Evaluation of the MORE-CARE wind power prediction platform. Perfrmance of the fuzzy logic based models | |
| CN114819337A (en) | Multi-task learning-based comprehensive energy system multi-load prediction method | |
| CN117713688B (en) | Low-efficiency identification and electric quantity lifting method of photovoltaic module under multi-orientation and inclination angles | |
| CN117709550B (en) | Energy consumption prediction method based on ARDL and convolutional neural network model | |
| CN117311295B (en) | Production quality improving method and system based on wireless network equipment |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210323 |