Solar radiation prediction system and abnormal operation estimation system

Abstract

To provide a solar radiation prediction system which predicts fully accurate solar radiation with less trouble and effort, and an abnormal operation estimation system capable of estimating absence or presence of occurrence of an abnormal operation of a solar panel in a state where the trouble and effort are reduced with the predicted solar radiation and full accuracy is obtained.SOLUTION: A solar radiation prediction system A1 predicts a solar radiation at an object position, and includes a server 1. The server 1 includes control means 8 and communication means 7. The communication means 7 is communicative with a solar radiation hour server HC capable of transmitting the solar radiation hours at a solar radiation hour providing position. The control means 8 obtains a solar radiation hours at a solar radiation hour providing position closest to the object position by controlling the communication means 7, and predicts the solar radiation at the object position on the basis of the solar radiation hours and the solar altitude at the object position. An estimation system E1 estimates an abnormal operation of a solar panel by using the solar radiation predicted by the solar radiation prediction system A1.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a solar radiation prediction system that predicts solar radiation, and an abnormal operation estimation system that estimates abnormal operation of a solar panel. [Background technology]

[0002] 2. Description of the Related Art A device disclosed in Japanese Patent Laid-Open Publication No. 2013-73537 (Patent Document 1) is known as a device for calculating an estimated value or a predicted value of the amount of power generated by photovoltaic power generation. This device calculates an estimate of the sunshine conditions from measured or predicted sunshine hours, and predicts the amount of power generated by photovoltaic power generation from the estimated sunshine conditions and the solar altitude. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2013-73537 A Summary of the Invention [Problem to be solved by the invention]

[0004] Although the above-mentioned device can provide an estimate of the sunshine conditions, it cannot provide a forecast of the amount of solar radiation, which can be used, for example, to estimate with sufficient accuracy the amount of power generated by a solar panel.

[0005] Therefore, a main object of the present invention is to provide a solar radiation prediction system that can predict solar radiation with sufficient accuracy and with little effort. Another main object of the present invention is to provide an abnormal operation estimation system that can estimate whether or not abnormal operation of a solar panel is occurring by using a predicted amount of solar radiation, reducing the amount of work required and achieving sufficient accuracy. [Means for solving the problem]

[0006] This specification discloses a solar radiation prediction system. The solar radiation prediction system may predict the amount of solar radiation at a target point. The solar radiation prediction system may include a server computer. The server computer may include a control means. The server computer may include a communication means. The communication means may be capable of communicating with a sunshine hour server HC capable of transmitting the sunshine hours at the sunshine hour providing point. The control means may control the communication means to obtain the sunshine hours at the sunshine hour providing point closest to the target point, or at the point closest to the target point among the sunshine hour selected points which are points selected from the sunshine hour providing points. The control means may predict the amount of solar radiation at the target point based on the sunshine hours and the solar altitude related to the target point.

[0007] This specification also discloses an abnormal operation estimation system. The abnormal operation estimation system may estimate abnormal operation of a solar panel using the solar radiation predicted by the solar radiation prediction system. The communication means may be capable of communicating with an aggregation server computer that aggregates values ​​of a smart meter, which is a power meter capable of communication. The communication means may be capable of communicating with an outside air temperature server computer that can transmit outside air temperature. The control means may control the communication means to obtain the value of the smart meter. The control means may control the communication means to obtain the outside air temperature. The control means may estimate an estimated power generation amount or power usage amount of an entity on which the solar panel is installed, or a related value that is a value related thereto, based on the value of the smart meter, the outside air temperature, and the solar radiation. The control means may extract a statistical outlier for the estimated power generation amount or power usage amount, or the related value. The control means may estimate abnormal operation of the solar panel based on the outlier value. Effect of the Invention

[0008] A main advantage of the present invention is that a solar radiation prediction system is provided that can predict solar radiation with sufficient accuracy and with little effort. Another main advantage of the present invention is that it provides an abnormal operation estimation system that can estimate whether or not abnormal operation of a solar panel is occurring by using the predicted amount of solar radiation, reducing the amount of work required and achieving sufficient accuracy. [Brief description of the drawings]

[0009] [Figure 1] 1 is an overall block diagram of an abnormal driving estimation system and related elements according to the present invention; [Diagram 2] 4 is a flowchart relating to an estimation program and an estimation application according to the present invention. [Diagram 3] FIG. 3A is a graph showing the relationship between actual measured values ​​of solar radiation and predicted values ​​of solar radiation, and FIG. 3B is a graph showing the relationship between actual measured values ​​of solar radiation and predicted values ​​of solar radiation when sunshine hours are not divided into predetermined time periods as boundary values. [Figure 4] FIG. 2 is a flowchart showing the calculation of the amount of electricity generated by solar panels when purchasing the entire amount. [Diagram 5] FIG. 2 is a flowchart for estimating abnormal operation of solar panels when purchasing the entire amount. [Figure 6] FIG. 6 is an explanatory diagram of a box plot relating to the extraction of outliers in FIG. 5 . [Figure 7] This is a graph showing the relationship between the outlier threshold value for the difference in power consumption calculated for each day within the past aggregation period over a specified one-year period and the judgment value, which is the difference in power consumption for that day, when a solar radiation prediction model related to sunshine hours and solar altitude for a main solar panel in the vicinity of Tsu, Mie Prefecture is divided into sections using specified times for sunshine hours as boundary values, and the graph shows the relationship when 15 sunshine hour selection points are selected in the solar radiation prediction. [Figure 8]This is a graph showing the relationship between the outlier threshold value for the difference in power energy calculated for each day within the past aggregation period over a specified one-year period and the judgment value, which is the difference in power energy for that day, in the case of a solar radiation prediction model related to sunshine hours and solar altitude for the solar panels of the same main entity in the vicinity of Tsu as in Figure 7, when a specified time for sunshine hours is not divided as a boundary value, and when 15 sunshine hour selection points are selected in the solar radiation prediction. [Figure 9] This is a graph showing the relationship between the outlier threshold value for the difference in power amount calculated for each day within the past aggregation period over a specified one-year period and the judgment value which is the difference in power amount for that day, in the case of a solar radiation prediction model related to sunshine hours and solar altitude not dividing the sunshine hours into specified times as boundary values, for the solar panels of the same main entity in the vicinity of Tsu as in Figure 7, and the graph shows the relationship between the outlier threshold value for the difference in power amount calculated for each day within the past aggregation period over a specified one-year period and the judgment value which is the difference in power amount for that day, in the case of a solar radiation prediction, in which the amount of solar radiation is obtained from only one of three solar radiation providing points. [Figure 10] FIG. 2 is a flowchart showing calculation of the amount of electricity used in surplus purchase. [Figure 11] FIG. 2 is a flowchart for estimating abnormal operation of solar panels during surplus purchase. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an embodiment of the present invention will be described together with modified examples with reference to the accompanying drawings. It should be noted that the embodiment is not limited to the following examples and modified examples.

[0011] FIG. 1 is an overall block diagram of an abnormal operation estimation system for a solar panel according to the present invention, ie, estimation system E1, and related elements. The estimation system E1 includes a solar radiation prediction system A1. The estimation system E1 uses a server 1 capable of reading and executing a solar panel abnormal operation estimation program (estimation program). The server 1 is a server computer. The server 1 may be at least one of a personal computer and a mobile terminal, or may be a combination of a plurality of devices appropriately connected via a network. The server computer of the solar radiation prediction system A1 and the server computer of the part of the estimation system E1 other than the solar radiation prediction system A1 may be separate. The solar radiation prediction system A1 uses a server 1 that can read and execute a solar radiation prediction program. The estimation program includes an insolation prediction program, which may be regarded as being independent of the estimation program.

[0012] The server 1 is installed in an electric power company. Here, the electric power company includes related organizations such as subsidiaries. The server 1 has a display means 2 for displaying information, an input means 4 for accepting input of information, a storage means 6 for storing information, a communication means 7 connectable to the leased line EL and the Internet IN, and a control means 8 for controlling these. For example, the display means 2 is a monitor, the input means 4 is at least one of a keyboard and a pointing device, the storage means 6 is at least one of a memory and a hard disk, the communication means 7 is a communication unit connectable to the leased line EL and the Internet IN, and the control means 8 is a CPU. The communication means 7 may be separated into a connection part for the dedicated line EL and a connection part for the Internet IN. When either the dedicated line EL or the Internet IN is not used, the unused part may be omitted.

[0013] The communication means 7 is connected to a tallying server CC (tallying server computer) installed in the electric power company via a dedicated line EL. The tallying server CC may be installed in another place or organization. The server 1 may be connected to the tallying server CC via the Internet IN. A plurality of smart meters SM are connected to the tallying server CC. The smart meters SM are power meters capable of communication. Here, each smart meter SM is connected to the tallying server CC via a wireless communication network. Note that each smart meter SM may be connected to the tallying server CC via the Internet IN or via a mobile phone network. Each smart meter SM is installed in an entity that has installed solar panels, in other words, an entity that has concluded an electricity purchase contract, and is a power meter capable of measuring the amount of electricity purchased for that entity, and further capable of calculating the amount of electricity purchased by time period by accumulating the amount of electricity purchased for each specified time period, and sending this to the aggregation server CC with entity identification information, i.e., entity ID. Here, the specified time period is a 30-minute time period, and there are 48 time periods per day. Furthermore, each amount of electricity purchased by time period with the entity ID attached is a so-called A route value. Note that multiple smart meters SM may be installed for one entity. Each entity sells a part or all of the electricity obtained by the solar panels to an electric power company. The solar panels can be regarded as a photovoltaic power generation facility. The solar panels are connected to a smart meter SM for each entity. The entities are, for example, power generation facilities. The smart meter SM is a power meter that is installed at an entity that generates electricity and can measure the amount of electricity purchased by the entity. The smart meter SM is connected to a solar panel via a power conditioner. The entity may also consume electricity. The scale of power generation (scale of the power generation facility) of the entity that is the subject of the abnormal operation estimation may be limited, for example, to a scale of contractual solar power generation capacity of 500 VA or less. A power generation facility with a contractual solar power generation capacity of 500 VA or less is called a small-scale power generation facility. The counting server CC has a database of purchased energy amounts that associates subject IDs, types of time periods (dates and times related to the time periods), and purchased energy amounts by time period.

[0014] There are two types of power purchase contracts between entities that have solar panels and electric power companies: full purchase and surplus purchase. A full-amount purchase is a contract in which the electric power company purchases all of the electricity generated by solar panels. In a full-amount purchase, if there is power consumption, the total amount of power consumed is sold separately by the electric power company according to how much is used. In this case, the smart meter SM obtains the amount of power generated by time period. The amount of generated power is a type of purchased power amount, which is the amount of power purchased by the electric power company. Surplus purchase is a contract in which electricity generated by solar panels is consumed by the home, the surplus electricity is purchased by the power company, and the shortage electricity is sold to the power company. In this case, the smart meter SM acquires the amount of surplus electricity by time period and the amount of shortage electricity by time period. The amount of surplus electricity is a type of purchased electricity amount. Moreover, the shortage electricity amount is also called the amount of electricity sold by the power company, because it is the amount of electricity sold to the main party. In surplus purchase, a distribution board is installed between the power conditioner and the smart meter SM. The distribution board is connected to the equipment that consumes electricity, such as electrical equipment, in the home, i.e., the main party to which the solar panels belong.

[0015] The communication means 7 is also connected to an outside temperature server TC (outside temperature server computer), a solar radiation server LC (solar radiation server computer), and a sunshine duration server HC (sunshine duration server computer) via the Internet IN. The outdoor temperature server TC is a server computer installed by the Japan Meteorological Agency, which is a meteorological observation organization. When a location and date are specified and an inquiry is made, the daily average outdoor temperature observed at the specified location and date is transmitted. The outdoor temperature may be the minimum or maximum temperature, a combination of these, or the weekly average temperature. The meteorological observation organization may be another organization other than the Japan Meteorological Agency. The solar radiation server LC is a server computer installed by the Japan Meteorological Agency, and when an inquiry is made by specifying a location and date, the solar radiation at the location and date is transmitted. Incidentally, instead of or together with the solar radiation, the slope solar radiation may be transmitted. When the slope solar radiation is transmitted, the inquiry may be made by specifying the angle of the slope. At least one of the slope solar radiation and the solar radiation may be calculated by the control means 8 according to the location and date or further the angle of the slope, or may be stored as a solar radiation database in the storage means 6. The solar radiation server LC may be integrated with the outdoor air temperature server TC, or may be installed in an organization different from the Japan Meteorological Agency. The sunshine hours server HC is a server computer installed by the Japan Meteorological Agency, and when an inquiry is made by specifying a location and date, the sunshine hours at the location and date are transmitted. The sunshine hours server HC is operated by the Japan Meteorological Agency, the same entity as the solar radiation server LC, and the sunshine hours server HC is associated with the solar radiation server LC. The sunshine hours may be calculated by the control means 8 according to the location, date, and weather information, or may be stored as a sunshine hours database in the storage means 6. At least two of the outside temperature server TC, the solar radiation server LC, the sunshine hours server HC, and the aggregation server CC may be a common server. The sunshine hours server HC may be integrated with at least one of the outside temperature server TC and the solar radiation server LC, or may be installed in an organization different from the Japan Meteorological Agency. Since devices that measure solar radiation are more complex and expensive than devices that measure sunshine hours, the number of locations related to the solar radiation server LC is smaller than the number of locations related to the sunshine hours server HC. For example, in the Tokai-Nagano region of Japan, solar radiation is provided at three locations, Nagoya, Shizuoka, and Nagano, while sunshine hours are provided at several dozen locations. Furthermore, at locations where solar radiation is provided, sunshine hours are also provided, and the solar radiation and sunshine hours are provided in association with each other. On the other hand, outside temperature is provided at approximately the same number of locations as sunshine hours. An information storage server computer (information storage server) that receives and stores at least one of the outside temperature, the amount of solar radiation, and the hours of sunlight at a predetermined timing and is separate from the server 1 may be installed at a power company, etc. In this case, the server 1 is capable of communicating with the information storage server and can obtain at least one of the outside temperature, the amount of solar radiation, and the hours of sunlight from the information storage server.

[0016] Furthermore, the communication means 7 is connected to a terminal T via the Internet IN. The terminal T is a mobile terminal in this embodiment, and like the server 1, has a terminal display means 12 for displaying information, a terminal input means 14 for accepting input of information, a terminal storage means 16 for storing information, a terminal communication means 17, and a terminal control means 18 for controlling these. The terminal display means 12 and the terminal input means 14 are, for example, displays with touch sensors, the terminal storage means 16 is, for example, a memory, the terminal communication means 17 is, for example, a communication unit connected to a mobile phone network so as to be able to perform data communications, and the terminal control means 18 is a CPU. The terminal storage means 16 of the terminal T stores an estimation application (estimation app, terminal side estimation program) that cooperates with the estimation program of the server 1, and the terminal control means 18 can execute the estimation app. The estimation app is associated with an entity that owns a solar panel, which is a user, and is associated with the smart meter SM of the entity, and holds an entity ID. Note that the entity ID in the terminal T and the entity ID in the server 1 may be different, and a correspondence table between them may be stored in at least one of the server 1 and the terminal T. At least one of the outdoor temperature server TC, the solar radiation server LC, the sunshine duration server HC, and the terminal T may be connected to the server 1 via a path other than the Internet IN, such as a dedicated line EL or an in-house communication. The various communication paths may include wired or wireless. The terminal T may be a PC (personal computer) placed on the user's body.

[0017] The solar radiation prediction program for the server 1 is stored in the storage means 6 and executed by the control means 8. The solar radiation prediction program has a function of predicting the solar radiation from the sunshine hours and the solar altitude. Here, the solar altitude is calculated by the server 1 based on the latitude of the location, as well as the date and time. The solar altitude may be obtained from a solar altitude server (solar altitude server computer), which is an external server that returns the solar altitude when the location, date and time are transmitted. The solar altitude server may be integrated with at least one of the outside air temperature server TC, the solar radiation server LC, and the sunshine hours server HC, or may be installed in an organization different from the Japan Meteorological Agency. The solar altitude is the meridian altitude here, but it may be the altitude at noon or at another time. The estimation program related to the server 1 is stored in the storage means 6 and executed by the control means 8. The estimation program includes a program for calculating the amount of power generated by the solar panels. The estimation application for the terminal T mainly transmits and receives various information to and from the server 1, and operates in conjunction with the solar radiation prediction program or the estimation program as appropriate. The power generation calculation program has a function of calculating the amount of power generated by the solar panels. The amount of power generated by the solar panels is one of the elements necessary for estimating abnormal operation of the solar panels. The power generation calculation program estimates the amount of power generated, which cannot be obtained directly from the smart meter SM and the aggregation server CC, based on a predetermined criterion. The predetermined criterion may be, for example, at least one function of the amount of solar radiation and the outside air temperature, or may be the sum of the predicted total amount of purchased power and the surplus power amount (-) or shortage power amount (+) actually measured by the smart meter SM. In addition, other programs may be added, any of the programs may be omitted, at least any of the programs may be divided into a plurality of programs, and at least any two of the programs may be integrated.

[0018] FIG. 2 is a flowchart of an estimation program and an estimation application that, when executed, performs the operation of estimating abnormal operation of a solar panel in the estimation system E1. Below, the operation when purchasing the entire amount will be mainly explained.

[0019] By executing the estimation application in the terminal T, the terminal control means 18 performs an initial setting process (step S1). If the initial setting has been completed, the control means 8 of the server 1 skips the initial setting related process in the estimation program based on step S1 and proceeds to step S2. In step S1, the terminal control means 18 accepts input of information related to the power generation capacity of the solar panels owned by the user (subject) and information related to the contract type, whether full purchase or surplus purchase. The input is performed, for example, by selecting from a pull-down list displayed on the terminal display means 12 using the terminal input means 14. When the user inputs all of this information (for example, when the input completion button is pressed after selections have been made in all of the pull-down lists), the terminal control means 18 stores each piece of information related to the initial settings (initial information) in the terminal storage means 16. The power generation capacity information may be the maximum amount of power generation, the average amount of power generation, the rated amount of power generation, or other values.

[0020] Next, in step S2, an estimation execution button is displayed, and when the user presses the estimation execution button using the terminal input means 14 of the terminal T, the terminal control means 18 transmits an execution command for the estimation program together with the subject ID to the server 1 via the terminal communication means 17. When the initial setting (step S1) has been performed, the initial information is also transmitted to the server 1.

[0021] When the control means 8 of the server 1 receives a command to execute the estimation program, the control means 8 executes the estimation program with the entity related to the entity ID as the target entity that is the subject of estimation, and according to the contract type, whether full purchase or surplus purchase. When the control means 8 receives each piece of information for the initial settings, it stores it in the storage means 6 in association with the entity ID. The control means 8 executes the estimation program to generally predict the amount of solar radiation using a solar radiation prediction program, and estimates abnormal operation of the solar panel of the subject based on the predicted amount of solar radiation, etc. The processing in the estimation program is repeatedly performed every predetermined period. That is, the processing in the estimation program is performed in units of a predetermined period. In this example, the predetermined period is one day. The server 1 may execute the estimation program based only on receipt of initial information, not based on an input to an estimation execution button on the terminal T. The predetermined period may also be specified on the terminal T. The repetition period of the solar radiation prediction program and the repetition period of the portion of the estimation program other than the solar radiation prediction program may be different from each other. Alternatively, the estimation program may be executed only once over a specific execution period (e.g., one month).

[0022] In the solar radiation prediction program, the control means 8 first requests the sunshine duration server HC to transmit the sunshine duration for the estimated target date for the location where the solar panel of the target subject is located, that is, the point corresponding to the installation position, that is, the point corresponding to the installation position. The control means 8 also calculates the solar altitude for that position or the point corresponding to the installation position (step S3). The installation position is the point where the solar radiation is predicted, and is the target point for the solar radiation prediction. The sunshine duration server HC finds the sunshine duration for the estimation target date and the point corresponding to the installation position, and transmits it to the server 1. The control means 8 of the server 1 causes the storage means 6 to store the sunshine duration received by the communication means 7.

[0023] The installation location corresponding point is a point, i.e., a sunshine hour selection point, selected in advance from the points for which sunshine hours are provided in the sunshine hour server HC, i.e., the sunshine hour providing points, and is the closest to the installation location. The purpose of selecting the sunshine hour selection point from the sunshine hour providing points is to reduce the amount of information processing while still achieving sufficient accuracy. However, if the amount of solar radiation is also provided at the nearest sunshine hour selection point, the control means 8 obtains the amount of solar radiation instead of the sunshine hour by transmitting the estimation target date and the installation point or the sunshine hour selection point to the solar radiation amount server LC. In this case, the amount of solar radiation is not estimated, the next step S4 is skipped, and the amount obtained from the solar radiation amount server LC is used. In addition, even if the amount of solar radiation is also provided at the sunshine hour selection point, the control means 8 may obtain the sunshine hour at the sunshine hour selection point and predict the amount of solar radiation in the next step S4. At least one of the predicted amount of solar radiation and the obtained amount of solar radiation may be a value obtained by correcting the amount of solar radiation on a horizontal plane to an azimuth angle of 0° and an inclination angle of 30°, or may be related to at least one of other azimuth angles and inclination angles, or may be something else. In addition, instead of the installation point, the latitude and longitude of the place (subject) where the solar panel is installed may be transmitted. Furthermore, when a solar altitude server is used, the solar altitude at the closest point among the solar altitude providing points, which are points where the solar altitude is provided in the solar altitude server, may be used as in the case of the sunshine hour, or the solar altitude at the closest point among the solar altitude selected points, which are points selected from the solar altitude providing points, may be used.

[0024] For example, 15 sunshine hour selection points are selected from several dozen sunshine hour provision points in the Tokai-Nagano region. The sunshine hour selection points are, for example, Nagano, Karuizawa, Matsumoto, Takayama, Suwa, Iida, Gifu, Nagoya, Shizuoka, Hamamatsu, Tahara, Yokkaichi, Tsu, Iga-Ueno, and Owase, taking into consideration at least one of the dispersion of the selection and the concentration of solar panels at the points. At least one of the sunshine hours and the amount of solar radiation may be obtained from the nearest sunshine hour provision point, not from the nearest sunshine hour selection point. If the installation location is in the vicinity of Hamamatsu, the closest sunshine hour selection point is Hamamatsu, and more accurate sunshine hours can be obtained at the installation location compared to the sunshine hours in Shizuoka, the nearest point where the amount of solar radiation is also provided. Similarly, if the installation location is in the vicinity of Owase, the closest sunshine hour selection point is Owase, and more accurate sunshine hours can be obtained at the installation location compared to the sunshine hours in Nagoya, the nearest point where the amount of solar radiation is also provided. On the other hand, if the installation location is near Nagoya, the closest sunshine duration selection point is Nagoya, and since the amount of solar radiation is also provided in Nagoya, a prediction of the amount of solar radiation based on the hours of sunshine, etc. is not made, and the amount of solar radiation for the estimated target date in Nagoya related to the solar radiation server LC is used.

[0025] Next, the control means 8 predicts the amount of solar radiation from the sunshine duration and solar altitude. The control means 8 constructs a prediction model in advance with the sunshine duration and solar altitude as explanatory variables, and predicts the amount of solar radiation at the position related to the subject ID on the day related to the processing by applying the sunshine duration and solar altitude obtained in step S3 to the prediction model (step S4). The prediction model here is a function with the sunshine duration and solar altitude as explanatory variables, and more specifically, is a two-variable linear function. The predicted amount of solar radiation is stored in the storage means 6. The prediction model may be a function of a different format, or may be a model of a machine learning system. The prediction model is constructed by dividing the case into a case where the sunshine hours are equal to or less than a predetermined time (here, 5 hours) and a case where the sunshine hours exceed the predetermined time. Of the prediction models, the form when the sunshine hours are below a predetermined time, i.e., a boundary value, is represented here by the following equation (1), and the form when the sunshine hours exceed the predetermined time, i.e., a boundary value, is represented here by the following equation (2). R a =1.9695×N+0.13624×α-2.2346 (1) R a =1.1520×N+0.26850×α-6.0236 (2) Here, R a is the amount of solar radiation, and R aThe unit is MJ / m 2 (Mega Joules per square meter). N is the unit of accumulated sunshine, and the unit of N is hours / day. α is the solar altitude, and the unit of α is degrees. The server 1 calculates the amount of insolation from the sunshine hours accumulated over a unit period (here, one day) and the solar altitude at noon. This calculation of the amount of insolation can reduce the amount of data and calculation while maintaining sufficient accuracy, compared to a calculation that accumulates the product of the sunshine hours for each specified hour multiplied by the sine of the solar altitude at that time. The specified time, i.e., the boundary value, may be other than 5 hours. Furthermore, the prediction model does not have to be constructed in such a manner, and may be constructed in three sections at the first specified time (first boundary value) and the second specified time (second boundary value), or may be constructed in four or more sections. Furthermore, the prediction model may be reconstructed using the sunshine duration and solar altitude from terminal T in the solar radiation prediction program while processing terminal T is being performed, or after processing terminal T is being performed. Furthermore, the solar altitude α may be in the form of a trigonometric function.

[0026] The prediction model was constructed as follows. That is, at the time of filing the present application, the solar radiation amount server LC provides solar radiation amounts at 35 locations in Japan, and 25 locations, which are a part of these locations, are used as learning data for determining a prediction model, and another 12 locations are used as test data for confirming the accuracy of the determined prediction model, thereby determining a prediction model. The 25 locations are Kumamoto, Matsuyama, Nagasaki, Nara, Matsue, Hikone, Kochi, Yamagata, Fukushima, Fukui, Oita, Miyazaki, Saga, Wakkanai, Maebashi, Morioka, Niigata, Kofu, Kagoshima, Akita, Utsunomiya, Abashiri, Muroran, Aomori, and Hakodate, and various locations with different climates, latitudes, and longitudes are selected. The 25 locations are Tokyo, Osaka, Sapporo, Sendai, Toyama, Hiroshima, Takamatsu, Fukuoka, Naha, Nagoya, Shizuoka, and Nagano, and various locations with different climates, latitudes, and longitudes are selected. At least one of the number of locations and the way of dividing the locations may be different from the above. At locations where the amount of solar radiation is provided, the sunshine duration server HC also provides the sunshine duration, and the solar altitude can also be calculated.

[0027] The control means 8 uses the learning data, which are the sunshine duration x (hours), the solar altitude y (°), and the solar radiation z (MJ / m 2 ) for which x≦5, plot a sufficient number m in the xyz space for each location and date. m is a natural number. For x≦5, plot x as x A Then, let y be y A Then, put z as z A Then, the control means 8 sets x A y A z A Each of the m plots in the space represents the explanatory variable x A ,y A The solar radiation function f A (x A ,y A ) where x A y A z A For each plot in space (x A1 ,y A1 ),(x A2 ,y A2 ),···,(x Am ,y Am ) for f A1 (x A1 ,y A1 ),f A2 (x A2 ,y A2 )···,f Am (x Am ,y Am ) and the actual solar radiation z A1 ,z A2 ,z Am and in f A1 (x A1 ,y A1 ) and z A1 Distance between,f A2 (x A2 ,y A2 ) and z A1 Distance between ,f Am (x Am ,y Am ) and z Amf that minimizes the sum of the distances between A (x A ,y A ), that is, by the least squares method in m plots, the function f A (x A ,y A ) is found. Note that the function f A (x A ,y A ) may be found by methods other than the least squares method, such as machine learning. Also, the control means 8 sequentially changes x, y, and z when x>5 to x B ,y B ,z B Then, x B y B z B By plotting n sufficient numbers of explanatory variables in the space, B ,y B The solar radiation function f for x>5 is expressed as B (x B ,y B ) as in the case where x≦5. That is, B y B z B For each plot in space (x B1 ,y B1 ),(x B2 ,y B2 ),···,(x Bn ,y Bn ) for f B1 (x B1 ,y B1 ),f B2 (x B2 ,y B2 )···,f Bn (x Bn ,y Bn ) and the actual solar radiation z B1 ,z B2 ,z Bn and in f B1 (x B1 ,y B1 ) and z B1 Distance between,f B2 (x B2 ,y B2 ) and z B1 Distance between ,fBn (x Bn ,y Bn ) and z Bn f that minimizes the sum of the distances between B (x B ,y B ) where n is a natural number.

[0028] Fig. 3A is a graph showing the relationship between the actual solar radiation amount and the predicted solar radiation amount for the test data. Fig. 3B is a graph showing the relationship between the actual solar radiation amount and the predicted solar radiation amount when a common prediction model is calculated using the sunshine hours x and the solar altitude y as explanatory variables without dividing the sunshine hours into a predetermined time, i.e., 5 hours, as a boundary value. The actual measured value of solar radiation is the solar radiation z for each location and date related to the test data. The predicted value of solar radiation is the sunshine duration x and solar altitude y for each location and date related to the test data multiplied by f determined from the learning data. A (x A ,y A ) or f B (x B ,y B ) is a value obtained by fitting the relationship between the sunshine hours x and the boundary value.

[0029] The control means 8 compares the actual measured value of the amount of solar radiation at the same location and on the same date with the predicted value, and if the absolute value of the difference falls within a predetermined range, it determines that the prediction model is sufficiently accurate. In Figures 3A and 3B, the horizontal axis is the predicted value of the amount of solar radiation, and the vertical axis is the actual measured value of the amount of solar radiation. The fact that the plot is located on the dashed line sloping upward to the right indicates that the predicted value of the amount of solar radiation matches the actual measured value, and the prediction related to the plot is perfect. In Figures 3A and 3B, the formula using x for y is a linear approximation formula for all plots, and R 2 is the coefficient of determination for the distribution of the plot. In both Figures 3A and 3B, the plots are ±5 MJ / m 2The results are within a range of about 0.9931 and 0.9839, respectively, in Figures 3A and 3B, which are very close to 1, the coefficient of y = x, which corresponds to the dashed line sloping upward to the right, and in this respect, too, the solar radiation prediction model is shown to be sufficiently accurate. Furthermore, in Figures 3A and 3B, the coefficients of the linear approximation equation are 0.9931 and 0.9839, respectively, which are very close to 1, the coefficient of y = x, which corresponds to the dashed line sloping upward to the right, and in this respect too, the solar radiation prediction model is shown to be sufficiently accurate. 2 is also close to 1, indicating that the solar radiation prediction model is sufficiently accurate in this respect as well. In addition, in FIG. 3A, the actual measured value of the solar radiation is 12 MJ / m compared to FIG. 3B. 2 In the region below about 100%, the number of plots in which the actual measured value is larger than the predicted value is reduced. Furthermore, in FIG. 3A, the coefficient of the linear approximation equation and the coefficient of determination R 2 are closer to 1. In at least some of these respects, the prediction model in the case of FIG. 3A can be said to be more accurate than the prediction model in the case of FIG. 3B.

[0030] Furthermore, the control means 8 calculates the error for the cases of Figs. 3A and 3B, and judges the accuracy of the prediction model of the amount of solar radiation based on the magnitude of the error. The errors here are RMSE (Root Mean Square Error) and MAPE (Mean Absolute Percentage Error). Note that the errors may be either RMSE or MAPE, or may be something other than RMSE or MAPE. The error is calculated by first finding the error for each location related to the test data, and then finding the average of the 12 locations. The error may also be calculated by other methods, such as by calculating the error for all locations at once. In the case of Figure 3B, the RMSE is 2.20 MJ / m 2 The MAPE was 0.176. In the case of Figure 3A, the RMSE was 1.82 MJ / m 2 The MAPE was 0.143, and both errors were smaller than those in Figure 3B. The control unit 8 may omit checking the accuracy of the prediction model using test data. The accuracy of the prediction model may be checked by either grasping using a graph or grasping using an error.

[0031] The control means 8 causes the storage means 6 to store the amount of solar radiation predicted by fitting the sunshine duration and solar altitude into the prediction model in association with the estimation target date and the subject ID. However, if the amount of solar radiation is also provided at the nearest sunshine hour selection point, the control means 8 stores the amount of solar radiation obtained from the solar radiation server LC in the memory means 6 in correspondence with the estimated target date and subject ID.

[0032] Then, the control means 8 estimates abnormal operation of the solar panels related to full purchase or surplus purchase based on the amount of solar radiation related to step S4. The control means 8 judges whether the total amount is to be purchased or the surplus amount is to be purchased (step S5), and if the total amount is to be purchased (Yes), the process proceeds to step S6, and if the surplus amount is to be purchased (No), the process proceeds to step S8. The control means 8 may apply a process for full purchase to surplus purchases that satisfy a predetermined condition, regarding it as full purchase. The predetermined condition here is, for example, a condition that the photovoltaic power generation ratio expressed as "contracted amount of power sold [kW] / contracted amount of power generated by the solar panel" is equal to or less than a predetermined value (for example, 1 / 20).

[0033] The occurrence of abnormal operation of the solar panels involved in full purchase in steps S6 and S7 is estimated based on a decrease in the amount of power generated at the time of processing compared to the ideal or standard amount of power generated statistically found, and is assumed to be abnormal or possibly abnormal in terms of power generation functions. Therefore, in full purchase, it is necessary to estimate the ideal or standard amount of power generated by the solar panels, i.e., the estimated amount of power generated. The estimated amount of power generated is calculated on a daily basis, with one day as the unit, and is statistically processed for a specified statistical period. Such daily estimated amount of power generated is called the estimated unit amount of power generated. The specified statistical period is three months here. In the case of full purchase, the estimated unit power generation amount is calculated by the following formulas (3) and (4) (step S6). Formulas (3) and (4) are included in the power generation amount calculation program. Equation (3) is the estimated unit solar power generation (solar power generation per day) per kW of generating capacity. In addition, on the right side of equation (3), the coefficient a of the amount of solar radiation may be other than 0.8944, the coefficient b of the outside air temperature may be other than -0.01670, and the intercept c may be other than 0.0154.

[0034] Estimated unit solar power generation [kWh / kW]= 0.8944× solar radiation [kWh / m 2 ]-0.01670 x Outside Temperature +0.0154 (3) Estimated unit power generation = Estimated unit solar power generation × Power generation capacity (4)

[0035] The amount of solar radiation in equation (3) is the predicted amount of solar radiation in step S4. The outside temperature in formula (3) is obtained from the outside temperature server TC and is the average temperature for one day. Note that a temperature other than the average temperature for one day may be used as the outside temperature. The power generation capacity in the formula (4) is the capacity of power generation in the solar panel, and is acquired as initial information. Note that the power generation capacity may be calculated from the power generation in the smart meter SM.

[0036] Equations (3) and (4) are based on the results of an analysis of more than 150 sample subjects related to the full-volume purchase. Assuming that the amount of generated electricity in the full-volume purchase is shown in equations (3) and (4), if the solar radiation data for the relevant location is used, the multiple correlation coefficient is 0.92, and therefore this assumption has considerable validity.

[0037] FIG. 4 is a flowchart showing the details of the calculation (step S6) of the estimated amount of power generated by the solar panels in the case of full purchase. The control means 8, which executes the power generation amount calculation program, refers to the solar radiation amount for that day in the storage means 6 (step S101), obtains the outdoor air temperature for that day from the outdoor air temperature server TC (step S102), and calculates the estimated unit power generation amount of the solar panel in the case of full purchase by applying the formulas (3) and (4) (step S103). Note that the order of steps S101 to S103 may be changed, and steps S101 to S103 may be performed simultaneously in parallel. The control means 8 stores the calculated estimated unit power generation amount of the solar panel in the case of full purchase in the storage means 6 for each day.

[0038] Next, the control means 8 estimates whether or not abnormal operation of the solar panels occurs during full purchase (step S7). As mentioned above, the suspected abnormal operation of solar panels is a decrease in their power generation function. In addition, the abnormal operation of the solar panels during the full purchase, which is the subject of the estimation, is not limited to a decrease in the power generation function. This also applies to the surplus purchase.

[0039] FIG. 5 is a flowchart showing details of step S7. In estimating whether or not abnormal operation of the solar panel occurs during full purchase, the control means 8 that executes the solar panel estimation program first performs a process of extracting the larger outlier from a set of values ​​obtained by subtracting the equivalent power generation amount equivalent to the actual power generation amount of that day from the estimated unit power generation amount for a predetermined statistical period (step S201). The equivalent power generation amount is expressed by the value of the smart meter SM in full purchase, and the control means 8 acquires the equivalent power generation amount of that day from the aggregation server CC. Here, the sum of the values ​​of the smart meter SM related to the time period from the specified start time (7:00) to the specified end time (16:00) is used as the equivalent power generation amount of that day. The value obtained by subtracting the equivalent power generation amount of that day from the estimated unit power generation amount, that is, the estimated power amount difference, is an ideal or standard value estimated from the amount of solar radiation, the outside temperature, and the power generation capacity, so the larger the estimated unit power generation amount is, the further the equivalent power generation amount of the solar panel of that day is from the ideal or standard, and the more the power generation function is deteriorated. The estimated power amount difference may be calculated in step S104 in accordance with the calculation of the estimated unit power generation amount. The equivalent power generation amount may be equal to the actual power generation amount of the solar panel, or may not be equal to the actual power generation amount of the solar panel, but may be equal to the actual power generation amount, although it is within a certain error range. Next, when the power amount difference at the processing point for estimating abnormal operation becomes larger than the larger outlier of the estimated power amount difference statistically found within a specified statistical period for a specified number of days or more within a specified period (Yes in step S202), the control means 8 estimates that the solar panel being fully purchased is operating abnormally (step S203). The predetermined period is the most recent seven days, and the predetermined number of days is three days. The processing time for estimating abnormal operation is the most recent day (the same day if the processing is performed at night, or yesterday if the processing is performed before night). That is, the control means 8 estimates that the solar panel in full purchase is operating abnormally if the power amount difference at the processing time, calculated by subtracting the equivalent power generation amount corresponding to the actual power generation amount of that day from the estimated power generation amount of that day, exceeds the larger outlier value related to the estimated power amount difference for three or more days out of the most recent seven days.

[0040] When the control means 8 estimates the occurrence of abnormal operation of the solar panel in the full purchase, it transmits abnormal operation information to the terminal T associated with the subject ID, and the terminal T performs display based on the received abnormal operation information by the estimation application. The user can recognize from the display that abnormal operation of the solar panel in the full purchase has occurred, and can take measures such as inspecting the solar panel. The terminal T accepts input of the subject ID in the estimation application and stores the subject ID in the terminal storage means 16, thereby associating the smart meter SM with the subject ID, which is an ID for distinguishing the smart meter SM from other smart meters SM. The abnormal operation information may be transmitted to the terminal T by email, or may be displayed by the terminal T accessing a dedicated website. The address of the email may be included in the initial information. Moreover, such processing for the terminal T may be omitted. The predetermined period may be six days or less or eight days or more, and the predetermined number of days may be two days or less or three days or more. The control means 8 may estimate abnormal operation when outliers appear for a certain number of consecutive days or more, such as when outliers appear for five consecutive days. At least one of the abnormal operation information and the display based thereon may indicate that the daily (estimated unit power generation amount - equivalent power generation amount) is statistically increasing. The predetermined statistical period may be other than three months. These examples of modifications regarding the predetermined number of days, the predetermined period, the specific number of days, the predetermined statistical period, and the abnormal operation information may be provided in other cases such as surplus purchase.

[0041] In step S201, statistical outliers may be extracted in any manner, but here they are extracted by the control means 8 creating a box plot.

[0042] 6 is an explanatory diagram of a box-and-whisker plot. Note that the control means 8 does not need to actually draw the box-and-whisker plot, and it is sufficient that a statistically equivalent process is performed. Alternatively, the control means 8 may display the box-and-whisker plot on the display means 2 or the like. The control means 8 creates a box-and-whisker plot for each subject ID. A box in the box-and-whisker plot is created for each day, with the horizontal axis of the box-and-whisker plot representing the date (day) of the estimated target date, and the vertical axis representing the amount of power (e.g., kWh (kilowatt-hours)) relating to the power amount difference. The horizontal axis of the box-and-whisker plot is not limited to a one-day width, but may be a half-day width or a two-day width, etc. The horizontal axis of the box-and-whisker plot may be set so that one box represents data over a predetermined statistical period or a portion thereof. The box-and-whisker plot may consist of only one box. The control means 8 calculates the first quartile Qt1 and the third quartile Qt3 for the difference in the amount of power (estimated unit power generation amount - equivalent power generation amount for that day) for each day, and the power amount difference that is larger than the third quartile Qt3 by more than 1.5 times the width (size of the box) of the power amount difference from the first quartile Qt1 to the third quartile Qt3 (outside the whiskers of 1.5 times the size of the box, i.e. the maximum value Max) is regarded as an outlier on the larger side. The maximum value Max is the largest value excluding the upper outlier, the lower outlier is the value obtained by subtracting 1.5 times the size of the box from the first quartile Qt1, and the minimum value Min is the smallest value excluding the lower outlier. The size of the box is also called the interquartile range (IQR), and is the third quartile Qt3 - the first quartile Qt1. In estimating abnormal operation of a solar panel, when the power amount difference, i.e. (estimated unit power generation amount - equivalent power generation amount for that day) is large, the equivalent power generation amount corresponding to the actual power generation amount deviates from the ideal power generation amount, so the upper outliers become a problem and are focused on. Note that the value of 1.5 times may be smaller or larger than 1.5. Also, at least one of the first quartile and the third quartile may be the first octile or the seventh octile, etc. In this way, since the power amount difference is statistically processed for each day within the specified statistical period, a certain amount of accumulation of the power amount difference is required, and the estimation is actually started three months after the specified statistical period is obtained. After three months, three months' worth of data is accumulated, and sufficient estimation can be performed. Note that, when the accuracy of the estimation is relatively important, a larger amount of data than three months may be used, such as a one-year amount of data being used for the specified statistical period. On the other hand, when the early start of the estimation is relatively important, a smaller amount of data than three months may be used, such as a two-week amount of data being used for the specified statistical period. The specified statistical period may be variable, such as two weeks at the start of the service and three months thereafter.

[0043] The control means 8 may find outliers by a method other than creating a box plot, such as creating a graph of a prediction interval or a statistically equivalent process. The control means 8 may also find outliers by using a combination of multiple methods. The prediction interval is an interval in which it is considered that there is an individual power amount difference for each statistical processing unit period (e.g., one day) with a certain probability, and for example, an interval with a 95% probability is set on the prediction interval graph. The horizontal and vertical axes of the prediction interval graph are the same as those of the box plot. The control means 8 then determines that the estimated power amount difference that is above the upper side of the prediction interval, i.e. (estimated unit power generation amount - equivalent power generation amount for that day), is an outlier on the larger side. The control means 8 may use a prediction interval with a probability of 99% or other probability. For reference, the upper boundary of the 95% prediction interval is shown in the following formula (5), and the lower boundary of the 95% prediction interval is shown in the following formula (6). In formulas (5) and (6), the t value is related to the t-test, and n is the number of samples (the number of plots of the power consumption difference). Upper bound of prediction interval = mean + 95% t-value × standard deviation × √(1+1 / n) (5) Lower prediction interval = mean - 95% t-value × standard deviation × √(1+1 / n) (6)

[0044] Figure 7 is a graph showing the relationship between the outlier threshold for the power amount difference, i.e., the estimated power amount difference, calculated for each day within the past aggregation period over the one-year period from April 1, 2021 to March 31, 2022, and the judgment value, which is the power amount difference for that day, i.e., the power amount difference at the time of processing, when the solar radiation prediction model related to sunshine hours and solar altitude for the main solar panels in the vicinity of Tsu is divided into when sunshine hours are below the boundary value and when sunshine hours exceed the boundary value, when 15 sunshine hour selection points are selected in the solar radiation prediction. Figure 8 is a graph showing the relationship between the outlier threshold for the difference in power volume calculated for each day within the past aggregation period over the one-year period from April 1, 2021 to March 31, 2022, and the judgment value, which is the difference in power volume for that day, when the solar radiation prediction model related to sunshine hours and solar altitude is unified rather than being separated into when sunshine hours are below the boundary value and when sunshine hours exceed the boundary value, for the solar panels of the same entity in the vicinity of Tsu as in Figure 7, when 15 sunshine hour selection points are selected in the solar radiation prediction. Figure 9 is a graph showing the relationship between the outlier threshold for the power amount difference calculated within the past aggregation period for each day over the one-year period from April 1, 2021 to March 31, 2022, and the judgment value, which is the power amount difference for that day, when the solar radiation prediction model for sunshine duration and solar altitude is unified rather than being separated into when sunshine duration is below the boundary value and when sunshine duration exceeds the boundary value, for the solar panels of the same entity in the vicinity of Tsu as in Figure 7, when the solar radiation prediction model is unified into when sunshine duration is below the boundary value and when sunshine duration exceeds the boundary value, and this graph shows the relationship when the solar radiation prediction is obtained only from one of the three solar radiation provision points in the solar radiation server LC. In Figures 7 to 9, the horizontal axis is the date, and the vertical axis is the power amount difference (kWh / day). When the power amount difference is a negative value, the equivalent power generation amount corresponding to the actual power generation amount is greater than the calculated estimated unit power generation amount. Therefore, in this case, it does not lead to the estimation of abnormal operation of the solar panel. 7 to 9, the past collection period, that is, the predetermined statistical period, is three months.

[0045] 7 to 9, the outlier threshold for the estimated power difference, which is the difference between the estimated unit power generation amount based on the solar radiation amount and the outside temperature constructed in the past collection period and the equivalent power generation amount, is a reference value for the outlier obtained statistically based on the estimated power difference for each day in the collection period for the solar panel related to the subject ID of the processing target. Here, the equivalent power generation amount for each day is the power generation amount equivalent to the actual power generation amount of the solar panel for the processing target day in the past collection period, and in the case of full purchase, it is the value of the smart meter SM and corresponds to the value obtained by accumulating the purchased power amount by time period within one day for the processing target day. The threshold value for the outliers related to the estimated power amount difference calculated within the past collection period is the value of the third quartile Qt3+(box size×1.5) of the estimated power amount difference on the box plot. The outlier threshold for the estimated power difference calculated within the past collection period can be considered to have a margin so that the result of subtracting the equivalent power generation amount equivalent to the actual power generation amount from the estimated unit power generation amount, which is the ideal power generation amount per day for the solar panel, may be larger to a certain extent (approximately the value of the third quartile Qt3 + (box size x 1.5) in the estimated power difference within the past collection period). Also, the outlier for the estimated power difference as a threshold can be considered to have a margin so that the result of subtracting the equivalent power generation amount equivalent to the actual power generation amount from the estimated power generation amount, which is the normally assumed power generation amount per day for the solar panel, may be larger to a certain extent (approximately the value of the third quartile Qt3 + (box size x 1.5) in the estimated power difference within the past collection period). 7 to 9, when the daily power amount difference at the processing target time point related to the abnormal operation estimation, i.e., the judgment value, exceeds such a large outlier calculated within the past aggregation period, it is treated as exceeding the threshold value, and serves as a basis for estimating the occurrence of abnormal operation of the solar panel. That is, when the daily actual power amount difference related to the abnormal operation estimation exceeds the outlier value as the threshold value for a predetermined number of days or more within a predetermined period, the control means 8 estimates the occurrence of abnormal operation of the solar panel. The daily power amount difference at the processing target time point is obtained by subtracting the equivalent power generation amount equivalent to the actual power generation amount of that day, i.e., the daily integrated value of the smart meter SM, from the estimated unit power generation amount calculated from the solar radiation and outside temperature of that day.

[0046] In fact, in the solar panels of the main solar panels near Tsu shown in Figures 7 to 9, there were no abnormal operations, i.e., no abnormalities in power generation functions due to a significant drop in the amount of power generated, during the one-year period from April 1, 2021 to March 31, 2022. In the case of FIG. 9, the amount of solar radiation used is that of Nagoya, which is the solar radiation providing point closest to Tsu. Therefore, due to differences in weather conditions, the amount of solar radiation in the vicinity of Tsu differs from that in Nagoya, which is used to calculate the estimated unit power generation amount, and the estimated unit power generation amount based on the amount of solar radiation, etc. differs from that expected in the vicinity of Tsu, and the outlier value of the power amount difference or the judgment value differs. As a result, there is a possibility that an erroneous estimation is made that the abnormal operation of the solar panel is estimated even though the abnormal operation of the solar panel does not actually occur. For example, in FIG. 9, there are situations in which the judgment value exceeds the threshold value in early January and late January, even though the solar panel is actually generating sufficient power with the solar radiation and outdoor temperature at that time, and there are situations in which abnormal operation may be estimated because the power generation is not commensurate with the estimated unit power generation amount based on the solar radiation and outdoor temperature at that time. This situation is thought to occur due to the discrepancy between the solar radiation in Nagoya and the solar radiation in Tsu during this period. The discrepancy between the amount of solar radiation in Nagoya and the amount of solar radiation in Tsu is, for example, due to the amount of solar radiation in Nagoya being considerably greater than the amount of solar radiation in Tsu on the same day, and due to the weather in Nagoya being sunny on that day while the weather in Tsu is cloudy, etc. Hereinafter, a situation in which the estimation system E1 estimates abnormal operation even though the solar panels are actually generating sufficient power may be referred to as a false abnormality estimation. In the case of FIG. 8, the amount of solar radiation is predicted based on the sunshine hours in Tsu, which is the sunshine hour selection point closest to the solar panels and is closer to the solar panels than Nagoya. Therefore, the deviation between the amount of solar radiation used to calculate the estimated unit power generation amount and the actual amount of solar radiation at the solar panel installation position is suppressed compared to the case of FIG. 9. Therefore, the occurrence of false anomaly estimation is suppressed compared to the case of FIG. 9. For example, in FIG. 8, the number of times the judgment value exceeds the threshold in early January and late January is reduced compared to FIG. 9. In the case of FIG. 7, the solar radiation amount is predicted based on the sunshine hours in Tsu, as in the case of FIG. 8. Furthermore, in the case of FIG. 7, the solar radiation prediction model is divided into sections with a predetermined time for the sunshine hours as a boundary value, so that the solar radiation amount is predicted more accurately than in the case of FIG. 8. Therefore, the occurrence of false anomaly estimation is further suppressed compared to the case of FIG. 8. For example, in FIG. 7, there are no situations in which the judgment value exceeds the threshold in early January and late January.

[0047] On the other hand, the estimation of abnormal operation of the solar panels during surplus purchase (steps S8 and S9) is based on the statistical increase in the new unit power consumption relative to the past unit power consumption (daily power consumption). The estimated abnormal operation of the solar panels is a decrease in the function related to power generation. The unit power consumption is also called self-consumption. That is, in the case of surplus purchase, if the actual amount of generated power decreases due to abnormal operation of the solar panels, the amount of surplus power (amount of purchased power) decreases or even disappears, and the amount of shortage power (amount of sold power), that is, the unit power consumption of the entity increases. Therefore, abnormal operation of solar panels in surplus purchase can be identified by the statistical increase in unit power consumption. In the case of surplus purchase, the daily unit electricity usage is calculated from the estimated unit electricity generation and the value of the smart meter SM on that day. The estimated unit power generation amount is calculated using the above formulas (3) and (4) in the case of full purchase. The unit power consumption is calculated by the following formula (7). Formula (7) is included in the power generation calculation program. Unit power consumption [kWh] = Estimated unit power generation + smart meter SM value for that day (7)

[0048] The value of the smart meter SM in formula (7) is the amount of power sold minus the amount of power purchased (from the power company's perspective), and is negative when there is surplus power (at the subject), and positive when there is a power shortage. Note that if the value of the smart meter SM is different from that described above and is positive when there is surplus power and negative when there is a power shortage, formula (7) may be changed to one that subtracts the second term from the first term. Similarly, formula (7) may be changed to another form that results in substantially the same calculation. Such a change to another form may be applied to other formulas. The smart meter SM value is the integrated value during the time period when solar power generation is possible, i.e., the time period from sunrise to sunset or a part of that time period. Here, the smart meter SM value is the integrated value during the time period from 7:00 to 16:00. The time period during which the integration of the smart meter SM value is limited may be changed depending on whether the total amount is purchased or the surplus amount is purchased, and also depending on the location, etc.

[0049] Equation (7) is based on the results of an analysis of over 150 sample entities (separate from the above sample entities) involved in surplus purchases. The multiple correlation coefficient for the assumption that the unit electricity usage in surplus purchases is shown in equation (7) is 0.8, and therefore this assumption is quite valid.

[0050] FIG. 10 is a flowchart showing details of the calculation of the unit electricity usage amount in surplus purchase (step S8). The control means 8, which executes the generated power amount calculation program, calculates the estimated unit generated power amount that is estimated as the daily generated power amount (S301 to S303), similarly to steps S101 to S103 in the case of full purchase.

[0051] Next, the control means 8 applies the obtained estimated unit power generation amount and the value of the smart meter SM on that day to equation (7) to calculate the unit power usage amount for surplus purchase (step S304). The control means 8 acquires the values ​​of the smart meter SM that belong to the above-mentioned time periods from the tallying server CC and adds them up.

[0052] Next, the control means 8 estimates abnormal operation of the solar panels in the surplus purchase (step S9). FIG. 11 is a flowchart showing details of step S9.

[0053] In this estimation, the control means 8 executing the solar panel estimation program first performs a process of excluding those unit power usage amounts accumulated during a predetermined statistical period that are smaller than a predetermined level (step S401). Here, the control means 8 excludes unit power consumption amounts that are twice the standard deviation σ of the unit power consumption amount, that is, 2σ or less (the lower 2σ), from the targets of calculation in the estimation. In this way, by excluding unit power consumption that is smaller than a certain level, cases where unit power consumption decreases without being due to abnormal operation of the solar panels, such as holidays, are excluded from the basic data for estimating abnormal operation, thereby improving the accuracy of the estimation. In addition, in certain cases, step S401 may not be executed, or step S401 may be omitted. Furthermore, the above-mentioned predetermined degree may be variously changed to 1.5σ, 3σ, or the like.

[0054] Next, the control means 8 performs a process of statistically extracting larger outliers from the estimated unit power consumption within a predetermined statistical period (step S402). Next, if the unit power consumption at the time being processed exceeds the larger outlier value as a threshold for a specified number of days or more within a specified period (Yes in step S403), the control means 8 presumes that the solar panel being purchased as surplus is operating abnormally (step S404). Here, the predetermined period is the most recent 14 days, and the predetermined number of days is 5 days. That is, if the unit power usage at the time of processing is an outlier on the larger side for 5 or more days out of the most recent 14 days, the control means 8 estimates abnormal operation of the solar panel in the surplus purchase.

[0055] The control means 8 performs the process of extracting outliers of unit electricity usage (step S402) in the same manner as in the case of full purchase. The extraction process may be performed using a box plot or a prediction interval. Here, the box plot is formed by one box for all unit power consumptions (excluding those excluded in step S401) within a predetermined period. At least one of the predetermined period and the predetermined number of days related to the outlier may be different for each type of entity at the time of full purchase and at least one of the time of surplus purchase. Furthermore, at the time of surplus purchase, abnormal operation of the solar panel may be estimated using related values, such as extracting statistical outliers using a unit power usage ratio, which is a value obtained by dividing unit power usage by a predetermined number (for example, average unit power usage), at the time of full purchase and at least one of the time of surplus purchase. The estimation system E1 may perform only the abnormal operation estimation in the case of full purchase, or may perform only the abnormal operation estimation in the case of surplus purchase. Furthermore, the estimation system E1 may also perform the abnormal operation estimation of other devices.

[0056] In the estimation system E1, even in the case of surplus purchase, the estimated unit power generation amount and ultimately unit power consumption are calculated more accurately based on the solar radiation, etc., more accurately predicted by the solar radiation prediction system A1, which uses the sunshine hours, etc. at sunshine hour acquisition points or sunshine hour selection points, which are more than the few solar radiation provision points, and abnormal operation of the solar panels is more accurately estimated, with false abnormality estimations suppressed.

[0057] The solar radiation prediction system A1 or estimation system E1 described above provides the following advantages. That is, the solar radiation prediction system A1 predicts the amount of solar radiation at a target point and includes a server 1. The server 1 includes a control means 8 and a communication means 7. The communication means 7 is capable of communicating with a sunshine hour server HC capable of transmitting sunshine hours at sunshine hour providing points. The control means 8 controls the communication means 7 to obtain the sunshine hours at the sunshine hour providing point closest to the target point, or at a point closest to the target point among sunshine hour selected points which are points selected from the sunshine hour providing points, and predicts the amount of solar radiation at the target point based on the sunshine hours and the solar altitude related to the target point. Therefore, the solar radiation prediction system A1 predicts the solar radiation with sufficient accuracy based on the solar radiation obtained from the sunshine hour providing point or the sunshine hour selecting point. Furthermore, the solar radiation prediction system A1 does not require the installation of a solar actinometer and a sunshine hour meter at the installation site of the solar panel when predicting the solar radiation, which reduces the effort required for prediction.

[0058] Furthermore, the communication means 7 can communicate with a solar altitude server capable of transmitting the solar altitude at a solar altitude providing point. The control means 8 controls the communication means 7 to obtain the solar altitude at the solar altitude providing point closest to the target point, or at the point closest to the target point among the solar altitude selected points which are points selected from the solar altitude providing points. Thus, the solar radiation prediction system A1 can predict the solar radiation with sufficient accuracy, further reducing the effort required for prediction. Alternatively, the control means 8 calculates the solar altitude from information related to the target point. Therefore, the solar radiation prediction system A1 can predict the solar radiation with a sufficient degree of accuracy, and the effort required for prediction is further reduced. The sunshine hour server HC is also associated with a solar radiation server LC that can transmit the amount of solar radiation at the solar radiation providing points. The number of sunshine hour providing points or the number of sunshine hour selection points is greater than the number of solar radiation providing points. Therefore, the solar radiation prediction system A1 can use data from points closer to the target point for which the amount of solar radiation is desired, compared to when the amount of solar radiation is obtained from a solar radiation server LC that has a smaller number of solar radiation providing points, and can predict the amount of solar radiation with sufficient accuracy.

[0059] Furthermore, the control means 8 predicts the amount of solar radiation at the target point based on a prediction model relating to the sunshine duration and the solar altitude. Therefore, the solar radiation prediction system A1 can predict the amount of solar radiation with sufficient accuracy, and the effort required for prediction is further reduced. In addition, the prediction model is divided into a plurality of categories based on a predetermined boundary value related to the sunshine duration. Therefore, in the solar radiation prediction system A1, the accuracy of the prediction model is improved, and the solar radiation can be predicted with sufficient accuracy.

[0060] In addition, the estimation system E1 estimates abnormal operation of the solar panel using the solar radiation predicted by the above-mentioned solar radiation prediction system A1. The communication means 7 can further communicate with an aggregation server CC that aggregates values ​​of a smart meter SM, which is a power meter capable of communication, and an outside air temperature server TC that can transmit outside air temperature. The control means 8 controls the communication means 7 to further obtain the value of the smart meter SM and the outside air temperature, estimate the estimated power generation amount or power usage of the entity on which the solar panel is installed, or related values ​​that are values ​​related thereto, based on the value of the smart meter SM, the outside air temperature, and the solar radiation, and further extracts statistical outliers for the estimated power generation amount or power usage, or related values, and estimates abnormal operation of the solar panel based on the outliers. Therefore, an estimation system E1 is provided that can estimate whether or not an abnormal operation of a solar panel has occurred, with reduced effort and with sufficient accuracy.

[0061] The control means 8 controls the communication means 7 to obtain the power generation capacity of the solar panel, and also obtains the value of the smart meter SM and the outside temperature, and estimates the estimated power generation amount of the solar panel based on the outside temperature, the amount of solar radiation, and the power generation capacity, and estimates that the solar panel is operating abnormally based on the estimated power difference obtained by subtracting the equivalent power generation amount from the estimated power generation amount, with a statistically obtained outlier as a threshold, based on the power amount difference at the time of processing exceeding the outlier, where the equivalent power generation amount corresponds to the actual power generation amount of the solar panel and is calculated from the value of the smart meter SM. Thus, an estimation system E1 is provided that can estimate the occurrence of abnormal operation of the solar panel while reducing the effort in full purchase and obtaining sufficient accuracy. On the other hand, the control means 8 controls the communication means 7 to obtain the power generation capacity of the solar panel, and also obtains the value of the smart meter SM and the outside temperature, and calculates the estimated power generation amount of the solar panel based on the outside temperature, the amount of solar radiation, and the power generation capacity, and further calculates the power consumption based on the estimated power generation amount and the value of the smart meter SM, and estimates that the solar panel is operating abnormally based on the power consumption at the time of processing exceeding the outlier value, using a statistically obtained outlier value as a threshold value. Thus, an estimation system E1 is provided that can estimate the occurrence of abnormal operation of the solar panel while reducing the effort in surplus purchase and obtaining sufficient accuracy.

[0062] In addition, the outliers are related to the box plot. Therefore, the outliers can be statistically extracted simply and accurately, and the accuracy of estimating abnormal operation of the solar panel is further improved. Furthermore, the estimation system E1 includes a terminal T capable of communicating with the server 1. The terminal T has a terminal display means 12 and is associated with an ID that distinguishes the smart meter SM connected to the solar panel. When the server 1 estimates that the solar panel is operating abnormally, the server 1 transmits abnormal operation information, which is information that the occurrence of the abnormal operation of the solar panel is estimated, to the terminal T associated with the ID corresponding to the estimated solar panel, via the communication means 7. The terminal T that receives the abnormal operation information performs a display based on the abnormal operation information on the terminal display means 12. Therefore, when the occurrence of the abnormal operation of the solar panel is estimated, the user who owns the terminal T can easily and quickly know that fact. [Explanation of symbols]

[0063] A1··solar radiation prediction system, E1··estimation system (abnormal operation estimation system), 1··server (server computer), 7··communication means, 8··control means, CC··aggregation server (aggregation server computer), HC··sunshine duration server (sunshine duration server computer), LC··solar radiation server (solar radiation server computer), SM··smart meter, T··terminal, TC··outdoor temperature server (outdoor temperature server computer).

Claims

1. A solar radiation prediction system for predicting solar radiation at a target point, A server computer is provided. The server computer includes a control means and a communication means, The communication means is capable of communicating with a sunshine hour server computer capable of transmitting the sunshine hour at the sunshine hour providing point, The control means Controlling the communication means to obtain the sunshine hours at the sunshine hours providing point closest to the target point or at a point closest to the target point among sunshine hours selected points which are points selected from the sunshine hours providing points; The amount of solar radiation at the target point is predicted based on the sunshine duration and the solar altitude at the target point. A solar radiation prediction system comprising:

2. Furthermore, the communication means is capable of communicating with a solar altitude server computer capable of transmitting the solar altitude at the solar altitude providing point, The control means The communication means is controlled to obtain the solar altitude at the solar altitude providing point that is closest to the target point, or at a solar altitude selected point that is a point selected from the solar altitude providing points and that is closest to the target point.

2. The solar radiation prediction system according to claim 1 .

3. The control means calculates the solar altitude from information related to the target point.

2. The solar radiation prediction system according to claim 1 .

4. The sunshine hour server computer is associated with a solar radiation amount server computer capable of transmitting the solar radiation amount at the solar radiation amount providing point, The number of the sunshine hour providing points or the number of the sunshine hour selecting points is greater than the number of the solar radiation providing points.

2. The solar radiation prediction system according to claim 1 .

5. The control means predicts the amount of solar radiation at the target point based on a prediction model relating to the sunshine duration and the solar altitude.

2. The solar radiation prediction system according to claim 1 .

6. The prediction model is divided into a plurality of categories according to a predetermined boundary value related to the sunshine hours.

6. The solar radiation prediction system according to claim 5.

7. 7. An abnormal operation estimation system that estimates abnormal operation of a solar panel by using the solar radiation predicted by the solar radiation prediction system according to any one of claims 1 to 6, The communication means is further capable of communicating with an aggregation server computer that aggregates values ​​of a smart meter, which is a power meter capable of communication, and an outside air temperature server computer that can transmit the outside air temperature; The control means Controlling the communication means to further obtain the value of the smart meter and the outside air temperature, Based on the value of the smart meter, the outside temperature, and the amount of solar radiation, an estimated amount of power generation or power usage of the entity in which the solar panel is installed, or a related value that is a value related thereto, is estimated; Further, a statistical outlier is extracted for the estimated power generation amount or the power consumption amount or the related value; An abnormal operation of the solar panel is estimated based on the outlier. An abnormal driving estimation system comprising:

8. The control means Controlling the communication means to obtain the power generation capacity of the solar panel, and to obtain the value of the smart meter and the outside air temperature; Estimating the estimated power generation amount of the solar panel based on the outside temperature, the amount of solar radiation, and the power generation capacity; Regarding the estimated power amount difference obtained by subtracting the equivalent power generation amount from the estimated power generation amount, the statistically obtained outlier is set as a threshold, and based on the power amount difference at the time of processing being greater than the outlier, it is estimated that the solar panel is operating abnormally; The equivalent power generation amount corresponds to the actual power generation amount of the solar panel and is calculated from the value of the smart meter. The abnormal driving estimation system according to claim 7 .

9. The control means Controlling the communication means to obtain the power generation capacity of the solar panel, and to obtain the value of the smart meter and the outside air temperature; Calculating the estimated power generation amount of the solar panel based on the outside temperature, the amount of solar radiation, and the power generation capacity, and further calculating the power usage amount based on the estimated power generation amount and the value of the smart meter; The outlier value obtained statistically for the amount of power usage is set as a threshold, and based on the amount of power usage at the time of processing exceeding the outlier value, it is estimated that the solar panel is operating abnormally. The abnormal driving estimation system according to claim 7 .

10. The outliers relate to box plots The abnormal driving estimation system according to claim 7 .

11. Further, a terminal capable of communicating with the server computer is provided, the terminal has a terminal display means and is associated with an ID that identifies the smart meter connected to the solar panel; When it is estimated that the solar panel is operating abnormally, the server computer transmits abnormal operation information, which is information indicating that the solar panel is operating abnormally, to the terminal associated with the ID corresponding to the estimated solar panel, via the communication means; The terminal that has received the abnormal operation information performs a display based on the abnormal operation information on the terminal display means. The abnormal driving estimation system according to claim 7 .

Images