KR20210074167A - Photovoltaic Generation Forecasting System - Google Patents

Abstract

The present invention relates to a solar power generation prediction system, which comprises a database, a machine learning module, and a prediction module. The database includes: a predictive weather database which receives and stores predicted weather information of a solar power plant from an organization which professionally predicts weather, including the Korea Meteorological Administration; a facility database which receives and stores facility information of the solar power plant from a manager terminal; an all-sky database which receives and stores all-sky information from an all-sky imaging device installed in the solar power plant; and a current weather database which receives and stores current weather information from sensors installed in the solar power plant. The machine learning module includes: an all-sky pattern learning module which generates all-sky pattern information by deriving a correlation between the all-sky information and the amount of power generation based on the all-sky information and the facility information; and a weather correction learning module for generating predicted weather correction information by correcting an error of the predicted weather information based on the predicted weather information and the current weather information. The power generation prediction module calculates the predicted power generation amount of the solar power plant for a future time point through the all-sky pattern information, the predicted weather correction information, and the facility information, and outputs the predicted power generation amount to a user terminal. The solar power generation prediction system can calculate a more accurate predicted power generation amount based on various information affecting the solar power generation.

Description

Photovoltaic Generation Forecasting System

The present invention relates to a solar power generation amount prediction system.

In general, solar power generation refers to a power generation method that converts light from the sun into electrical energy by using the photovoltaic effect in which an electromotive force is generated at an interface between a semiconductor or an electrolyte solution by irradiation of light.

Unlike conventional energy sources such as fossil raw materials, solar power generation is a clean energy source without risks such as greenhouse gas emissions, noise, and environmental destruction that cause global warming, and there is no concern about exhaustion.

In addition, solar power generation has the advantage of being relatively free of restrictions on installation and low maintenance cost compared to other renewable energy such as wind power or sea water power.

However, since such photovoltaic power generation has a large output variation according to weather and has a disadvantage in that efficiency is reduced according to temperature, generation prediction is a very important technology for stable operation of the power system and continuous supply of power generation.

That is, if the exact amount of power generation of the solar power plant is predicted, the power transaction in the power market and the planning and operation of the power system can be made more easily.

Accordingly, in recent years, various technologies for predicting the amount of power generation of a solar power plant have been developed.

As a prior art for this, there is a method of predicting the amount of power generation of a solar power generator disclosed in Korean Patent Application Laid-Open No. 10-2019-0131808 (published on November 27, 2019).

As shown in FIG. 1, the conventional apparatus and method for predicting the amount of power generation of a solar power generator predicts the power generation amount through the meteorological information collected from the Meteorological Administration, etc. There is this.

More specifically, the conventional method of predicting the amount of power generation of a solar power generator includes the steps of collecting weather information necessary to predict the amount of power generation of the photovoltaic generator, and a measurement time point at which the power generation amount of the photovoltaic generator is predicted using the collected weather information. calculating the temperature change rate in , determining the predicted temperature at the measurement time by using the calculated temperature change rate, and confirming the amount of insolation at the measurement time using the predicted temperature at the determined measurement time and predicting the amount of power generation of the photovoltaic generator at the measurement time by using the solar radiation at the confirmed measurement time.

That is, the conventional method for predicting the amount of power generation of a solar power generator calculates a temperature change rate through weather information and predicts the temperature through this to predict the amount of power generation.

Accordingly, the conventional method of predicting the amount of power generation of a solar power generator has a problem in that the accuracy is lowered because other conditions depending on the amount of cloud or equipment are not reflected.

Republic of Korea Patent Publication No. 10-2019-0131808 (published on November 27, 2019)

The present invention has been devised to solve the above problems, and provides a system for predicting the amount of solar power generation that can predict the amount of power generation of a solar power plant for a future point in time through weather information, facility information, and all-sky photos collected at the site. will be.

In addition, the present invention is to provide a system for predicting the amount of solar power generation in which accuracy can be gradually improved by applying a machine learning module through big data such as weather, equipment, and all-weather.

The objects of the present invention are not limited to the above-mentioned objects, and other objects and advantages of the present invention not mentioned may be understood by the following description, and will be more clearly understood by the embodiments of the present invention.

Moreover, it will be readily apparent that the objects and advantages of the present invention may be realized by the means and combinations thereof indicated in the claims.

In order to solve the above problems, the present invention provides a solar power generation amount prediction system including a database, a machine learning module, and a generation amount prediction module.

More specifically, the database includes a forecast weather database, a facility database, an all-weather database, and a current weather database.

The predictive weather database receives and stores forecasted weather information of the solar power plant from an organization that professionally predicts the weather, including the Korea Meteorological Administration.

In addition, the facility database receives and stores facility information of the solar power plant from the manager terminal.

More preferably, the facility database may include a facility profile for the amount of power generated according to the latitude, longitude, photovoltaic panel, and each information about the inverter of the photovoltaic power plant.

In addition, the all-sky database receives and stores all-sky information from an all-sky imaging device installed in the solar power plant.

In addition, the current weather database receives and stores current weather information from sensors installed in the solar power plant.

On the other hand, the machine learning module is based on the all-season information and the facility information, and the all-season pattern learning module that generates the all-season pattern information by deriving the correlation between the all-season information and the amount of power generation, and the predicted weather information and the current weather information. and a weather correction learning module configured to correct an error of the predicted weather information to generate predicted weather correction information.

Meanwhile, the predicted weather information and the current weather information may include information on temperature, cloudiness, wind speed, wind direction, insolation, humidity, precipitation, and fine dust, respectively.

In addition, the facility information may include information on the amount of power generation, latitude, longitude, solar panel, and inverter.

In addition, the all-sky information may include information on the type of cloud, the color of the cloud, the amount of the cloud, and the color of the sky.

More specifically, the machine learning module selects an input variable and an output variable, respectively, and selects an input variable having high correlation with the output variable among the selected input/output variables through importance analysis to configure a data set of input/output variables, At least one artificial neural network model for each cluster having similarity within the data set can be generated through cluster analysis of the data set of variables.

More preferably, the all-season pattern learning module includes each information on the type of cloud, the color of the cloud, and the color of the sky as input variables, and the output variable includes information on the amount of power generation, so that the output of the artificial neural network model is all-natural. It may be pattern information.

In addition, the weather correction learning module includes information on the error between the predicted weather information and the current weather information as input variables, and the output variables are future temperature, cloudiness, wind speed, wind direction, insolation, humidity, precipitation, fine dust. The output of the artificial neural network model including information about the prediction may be prediction weather correction information.

On the other hand, the power generation prediction module may calculate the predicted power generation amount of the solar power plant for a future time through the all-sky pattern information, the predicted weather correction information, and the facility information, and output it to the user terminal.

More preferably, the power generation prediction module may calculate the predicted power generation amount of the solar power plant for a future time through at least one or more of the artificial neural network model and facility profile.

Also, the generation amount prediction module may output the calculated predicted generation amount to the user terminal for each preset time period.

The solar power generation prediction system according to the present invention can calculate a more accurate predicted power generation amount based on various information affecting the solar power generation.

In addition, the solar power generation amount prediction system according to the present invention gradually improves the accuracy of the predicted generation amount by applying a machine learning module using big data including meteorological, facility, and all-weather information. to make operations easier.

In addition to the above-described effects, the specific effects of the present invention will be described together while describing specific details for carrying out the invention below.

1 is a diagram illustrating a conventional power generation amount prediction device.
2 is a diagram illustrating a flow of information for constructing a database in the solar power generation amount prediction system according to an embodiment of the present invention.
3 is a diagram illustrating a solar power generation amount prediction system according to an embodiment of the present invention.
4A and 4B are diagrams illustrating information flows for constructing an all-season information learning module and a weather information learning module in the solar power generation amount prediction system according to an embodiment of the present invention, respectively.
5 is a diagram illustrating a flow of information for constructing a power generation prediction module in a solar power generation amount prediction system according to an embodiment of the present invention.

The above-described objects, features and advantages will be described below in detail with reference to the accompanying drawings, and accordingly, those of ordinary skill in the art to which the present invention pertains will be able to easily implement the technical idea of the present invention. In describing the present invention, if it is determined that a detailed description of a known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description will be omitted. Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same reference numerals are used to indicate the same or similar components.

In the following, that an arbitrary component is disposed on the "upper (or lower)" of a component or "upper (or below)" of a component means that any component is disposed in contact with the upper surface (or lower surface) of the component. Furthermore, it may mean that other components may be interposed between the component and any component disposed on (or under) the component.

Also, when it is described that a component is "connected", "coupled" or "connected" to another component, the components may be directly connected or connected to each other, but other components are "interposed" between each component. It is to be understood that “or, each component may be “connected,” “coupled,” or “connected” through another component.

As used herein, the singular expression includes the plural expression unless the context clearly dictates otherwise.

In the present application, terms such as "consisting of" or "comprising" should not be construed as necessarily including all of the various components or various steps described in the specification, some of which components or some steps are It should be construed that it may not include, or may further include additional components or steps.

Throughout the specification, when “A and/or B” is used, it means A, B or A and B, unless specifically stated to the contrary, and when “C to D” is used, it means that there is no specific contrary description. Unless otherwise specified, it means C or more and D or less.

Hereinafter, a solar power generation amount prediction system and method according to some embodiments of the present invention will be described.

2 is a diagram illustrating a flow of information for constructing a database in a system for predicting solar power generation according to an embodiment of the present invention.

And, FIG. 3 is a diagram illustrating a solar power generation amount prediction system according to an embodiment of the present invention.

4A and 4B are diagrams illustrating information flows for constructing an all-sky information learning module and a weather information learning module in the solar power generation amount prediction system according to an embodiment of the present invention, respectively.

5 is a diagram illustrating a flow of information for constructing a power generation prediction module in the solar power generation amount prediction system according to an embodiment of the present invention.

2 to 5 , the solar power generation amount prediction system according to the present invention includes a database 100 , a machine learning module 200 , and a generation amount prediction module 300 .

More specifically, the database 100 includes a forecast weather database 110 , a facility database 120 , an all-season database 130 , and a current weather database 140 .

The predicted weather database 100 receives and stores the predicted weather information of the solar power plant P from an organization that professionally predicts the weather including the Meteorological Agency C in real time or at preset time intervals.

In this case, the predicted weather information may include information on future temperature, cloudiness, wind speed, wind direction, insolation, humidity, precipitation, fine dust, and the like.

More preferably, the predicted weather database 100 may store predicted weather information for each location information based on location information of a plurality of regions where the solar power plant P is installed.

Meanwhile, the facility database 120 receives and stores facility information of the solar power plant P from the manager terminal T1 in real time or at preset time intervals. The database 100 may include a receiver (not shown) capable of receiving predicted weather information, facility information, all-weather information, and current weather information. In addition, the database 100 includes data or information stored in each database (eg, the predicted weather database 110 , the facility database 120 , the all-season database 130 , and the current weather database 140 ) into a machine learning module ( 200), a power generation prediction module 300, and a transmitter (not shown) for transmitting to at least one of the manager terminal T1. For example, the facility database 120 may include the manager terminal T1. It is possible to receive and store facility information of the solar power plant P from the receiving unit (not shown).The facility database 120 receives facility information obtained through an application installed in the manager terminal T1 from the receiving unit. (not shown) can be received periodically or in real time.The facility information may include information on the current state of the facility of the solar power plant P. The facility information includes the current power generation facility capacity, Information on current supply capacity, current maximum power, current reserve power, current supply reserve ratio, current average power, current generation amount, current fuel consumption, current thermal efficiency, current power consumption rate, current frequency maintenance rate, and current failure The manager terminal T1 may periodically check the status of the facility of the solar power plant P through the application, such as whether there is a failure, and transmit facility information according to the check result to the facility database 120 . In addition, the manager terminal T1 may communicate in real time with the user terminal T2 through the application The manager terminal T1 transmits the facility information in real time through the application to the user terminal ( T2), the user of the user terminal T2 can not only check the power generation forecast amount transmitted through the generation amount prediction model 300, but also use the facility information received from the manager terminal T1. It is possible to check the current state of the facilities of the photovoltaic power plant P, and whether there is a failure.

The photovoltaic power plant P may obtain a measurement value corresponding to each facility information from the corresponding equipment of the photovoltaic power plant P, and compare the obtained measurement value with a threshold value. For example, when the acquired measurement value exceeds the threshold value, the solar power plant P transmits the acquired measurement value to the database 100 (eg, equipment database 120), or an administrator terminal (T1) can be transmitted. Alternatively, the solar power plant P may transmit the facility information to the database 100 (eg, the facility database 120 ) or to the manager terminal T1 .

The manager terminal T1 may receive various information (eg, facility information) on facilities of the photovoltaic power plant P from the photovoltaic power plant P through an installed application. The solar power plant P may periodically transmit facility information to the manager terminal T1 or the user terminal T2 through self-diagnosis. The manager terminal T1 may receive and store facility information transmitted from the solar power plant P through the application. In addition, the manager terminal T1 may diagnose whether the equipment of the solar power plant P has a failure through the application. In addition, the user terminal T2 receives the facility information on the facility of the photovoltaic power plant P through the pre-installed application from the photovoltaic power plant P or by receiving it from the facility database 120 , the photovoltaic It can be determined whether there is a failure in the equipment of the power plant (P).

In addition, the manager terminal T1 may receive various information (eg, facility information) about the facilities of the solar power plant P from the manager through the application. The manager terminal T1 periodically checks the presence or absence of abnormalities in the facilities of the photovoltaic power plant P through the application, and based on the check, a result value for the facilities of the photovoltaic power plant P (eg: information, data, etc.) may be transmitted to the facility database 120 .

In addition, the facility database 120 may obtain a measurement value corresponding to each facility information from the corresponding equipment of the photovoltaic power plant P, and compare the obtained measurement value with a threshold value. For example, when the acquired measurement value exceeds the threshold value, the facility database 120 may transmit the acquired measurement value to the user terminal T2 and/or the manager terminal T1 . At least one of the user terminal T2 and the manager terminal T1 may display a result of the obtained measurement value by voice or a pop-up.

In this case, the manager terminal T1 is a terminal including a PC, a mobile device, etc. of the manager of the solar power plant P or the manager of the solar power generation amount prediction system according to the present invention, and the facility database 120 and the wired/wireless communication network can be connected to

In addition, the facility information may include information on the amount of power generation, latitude, longitude, solar panel, and inverter.

More specifically, the information on the amount of power generation means the amount of power actually produced by the photovoltaic power plant (P).

In addition, the latitude and longitude include information on the location where the solar panel of the solar power plant (P) is installed.

In addition, the information on the solar panel and the inverter may include the respective efficiency, the efficiency according to the combination of the solar panel and the inverter, the array state of the solar panel, the period of use, the production date, and the like.

More preferably, the facility database 120 may include a facility profile for the amount of power generated according to the latitude, longitude, photovoltaic panel, and each information about the inverter of the photovoltaic power plant.

In addition, the all-sky database 130 receives and stores all-sky information from the all-sky imaging device (S) installed in the solar power plant (P) in real time or at preset time intervals.

Accordingly, the all-sky imaging device (S) may be connected to the all-sky database 130 through a wired/wireless communication network.

In addition, the all-sky information includes cloud types such as cirrus, cirrocumulus, cirrostratus, altocumulus, stratocumulus, cumulus, stratocumulus, stratus, stratostratus, cumulonimbus, etc., cloud color (saturation contrast), cloud amount, sky color (saturation, contrast) It may be a photo or an image including information about it.

In addition, the current weather database 140 receives and stores the current weather information from the sensors (S) installed in the solar power plant (P) in real time or at preset time intervals.

The current weather information may include information on temperature, cloudiness, wind speed, wind direction, insolation, humidity, precipitation, fine dust, and the like at the present time.

In addition, the sensor (S) is preferably composed of an electronic sensor that can measure the current weather information in real time or every preset time interval.

That is, there is a difference that the predicted weather information is location-based prediction information, whereas the current weather information is actual measurement information of the current time of the area in which the solar power plant P is located.

More preferably, in the database 100 of the solar power generation amount prediction system according to the present invention as described above, the machine learning module 200 and the generation amount prediction module 300 may be connected to each other through a wired/wireless communication network including the Internet.

On the other hand, the machine learning module 200 includes an all-season pattern learning module 210 that generates all-season pattern information by deriving a correlation between all-season information and power generation based on the all-season information and the facility information, and the predicted weather information and the It may include a weather correction learning module 220 for generating predicted weather correction information by correcting an error of the predicted weather information based on current weather information.

More specifically, the machine learning module 200 selects an input variable and an output variable, respectively, and selects an input variable having a high correlation with an output variable among the selected input/output variables through importance analysis to configure a data set of input/output variables, , it is possible to generate at least one artificial neural network model for each cluster having similarity within the data set through cluster analysis of the data set of the input/output variable.

In addition, the all-sky pattern learning module 210 includes each information on the type of cloud, the color of the cloud, and the color of the sky as input variables, and the output variable includes information on the amount of power generation to output the artificial neural network model. may be all-world pattern information.

For example, the omni-fat pattern learning module 210 may generate omni-fatal pattern information in the following way.

In the visible region, the short-wavelength blue series is blue because its scattering is stronger than the long-wavelength red series, and the clouds appear white because they are scattered for all wavelengths, and have a darker color depending on the thickness of the clouds.

Accordingly, in the picture of the area where there are no clouds, the shorter the wavelength, the higher the average brightness in the order of B (Blue), G (Green), R (Red), and the closer the zenith angle is to noon, the lower the amount of incident solar radiation. As the increase increases, the scattering becomes stronger, and the average brightness of both RGB increases.

On the other hand, the average brightness of RGB is similarly distributed in the photo of the area where clouds exist. Therefore, it is possible to distinguish the sky and cloud pixels in the image by the RGB contrast ratio.

In particular, since the mean standard deviation of G (Green) is the smallest, so there is no large deviation for each pixel, and forms a relatively stable value, the amount of cloud cover can be calculated according to the frequency distribution of GBR.

Also, more preferably, the all-sky pattern learning module 210 may analyze the amount of insolation according to the type of cloud, cloud saturation, sky saturation, contrast, and the like.

Accordingly, the all-season pattern learning module 210 may generate the all-season pattern information by patterning the correlation between the amount of power generation with respect to the amount of clouds and the type of location, respectively.

On the other hand, the weather correction learning module 220 includes information on the error between the predicted weather information and the current weather information as input variables, and the output variables are temperature, cloudiness, wind speed, wind direction, insolation, humidity, precipitation, The output of the artificial neural network model including information on fine dust may be predictive weather correction information.

That is, the weather correction learning module 220 uses the predicted weather information based on the location and the current weather information based on the actual measurement as described above to more accurately affect the solar power generation amount of the solar power plant (P). Weather information can be predicted.

On the other hand, the power generation prediction module 300 calculates the predicted power generation amount of the solar power plant for a future time through the all-sky pattern information, the predicted weather correction information, and the facility information, and outputs it to the user terminal T2. can

Accordingly, the solar power generation amount prediction system according to the present invention can calculate a more accurate predicted power generation amount based on various information affecting the solar power generation.

In addition, the user terminal T2 may include a PC, a mobile device, a server, etc. of an individual or group who want to read the predicted power generation amount of the solar power plant.

More specifically, the power generation prediction module 300 is the solar power plant for a future time through at least one artificial neural network model generated by the machine learning module 300 and the facility profile received from the facility database 120 . The predicted power generation of (P) can be calculated.

That is, the generation amount prediction module 300 calculates the predicted generation amount of the solar power plant P through the facility profile, all-weather pattern information, and predicted weather correction information, as shown in FIG. 5 ,

Accordingly, in the solar power generation prediction system according to the present invention, by applying a machine learning module using big data including weather, facility, and all-weather information, the accuracy of the predicted generation amount is gradually improved, so that To facilitate planning and operation.

Also, the generation amount prediction module 300 may output the calculated predicted generation amount to the user terminal T2 for each preset time period.

In addition, the solar power generation amount prediction system according to the present invention may be formed so that the database 100 can be built through information input through an application built into the user terminal T2.

In addition, the solar power generation prediction system according to the present invention can accurately calculate the predicted generation amount when clear facility information is provided regardless of the size of the solar power plant P for which the predicted generation amount is to be calculated.

Accordingly, the solar power generation prediction system according to the present invention can not only calculate the predicted power generation amount of the previously built solar power plant P, but also calculate the predicted power generation amount of the solar power plant P to be built in the future. Therefore, it can be easily used in the solar power generation business.

As described above, the present invention has been described with reference to the illustrated drawings, but the present invention is not limited by the embodiments and drawings disclosed in this specification, and various methods can be obtained by those skilled in the art within the scope of the technical spirit of the present invention. It is obvious that variations can be made. In addition, although the effects according to the configuration of the present invention are not explicitly described and described while describing the embodiments of the present invention, it is natural that the effects predictable by the configuration should also be recognized.

W. Meteorological Agency
P. Solar power plant
C. All-sky imaging device
S. sensor
T1. admin terminal
T2. user terminal
100. Database
110. Predictive Weather Database
120. Equipment database
130. The Whole World Database
140. Current Weather Database
200. Machine Learning Module
210. All-world pattern learning module
220. Weather correction learning module
300. Power Generation Prediction Module

Claims (2)

a forecast meteorological database that receives and stores forecasted meteorological information of places where solar power plants are installed from professional forecasting agencies, including the Meteorological Administration;
a facility database for receiving and storing facility information indicating the current state of the facility of the solar power plant from the manager terminal;
an all-sky database for receiving and storing all-sky information from an all-sky imaging device installed in the solar power plant; and
a database including a current weather database for receiving and storing current weather information for a current point in time from sensors installed in the solar power plant;
The input variable is the facility information and the all-season information, and the output variable is all-cheon pattern information that is generated through cluster analysis on a data set composed of information on the amount of power generation of the solar power plant and patterned the correlation between the all-heaven information and the amount of power generation. An all-world pattern learning module that is an artificial neural network model that outputs
The input variable is information on the error between the predicted weather information and the current weather information, and the output variable is an artificial neural network model that is generated through cluster analysis on the data set composed of the predicted weather information and outputs predicted weather correction information. a machine learning module including a correction learning module; and
A power generation prediction module for calculating the predicted power generation amount of the solar power plant for a future point in time through the all-sky pattern information and the predicted weather correction information, and outputting it to a user terminal at a user request or at a preset time,
The facility information is
It is obtained by checking the current state of the solar power plant through the application installed in the manager terminal,
Current power generation facility capacity, current supply capacity, current maximum power, current reserve power, current supply reserve ratio, current average power, current generation amount, current thermal efficiency, current power consumption rate, current frequency maintenance rate, and current for the facilities of the solar power plant Solar power generation prediction system including information on the presence or absence of failure.
According to claim 1,
The facility database is
Compare each of the measured values for the facility information with a threshold value,
A solar power generation prediction system for transmitting the measured value exceeding the threshold to the manager terminal and the user terminal through a transmitter.

Images