Photovoltaic power generation management device, photovoltaic power generation management method, and photovoltaic power generation management program

Abstract

To provide a photovoltaic power generation management device, a photovoltaic power generation management method, and a photovoltaic power generation management program that are capable of predicting the snow accretion condition on a photovoltaic (PV) panel with high accuracy.SOLUTION: According to one embodiment, a photovoltaic power generation management device includes an acquisition unit that acquires photographed image data of a predetermined area including an installation location of a photovoltaic (PV) panel at a reference time point and weather forecast data at a prediction target time point. The photovoltaic power generation management device further includes a snow accumulation prediction unit that predicts the snow accumulation condition of the predetermined area at the prediction target time point based on the photographed image data and weather forecast data acquired by the acquisition unit. The photovoltaic power generation management device further includes a snow accretion prediction unit that predicts the snow accretion condition on the PV panel at the prediction target time point based on the prediction result of the snow accumulation condition. The acquisition unit further checks whether there is a missing part in the weather forecast data and acquires weather forecast data for a timing prior to the prediction target time point.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] An embodiment of the present invention relates to a photovoltaic power generation management device, a photovoltaic power generation management method, and a photovoltaic power generation management program. [Background technology]

[0002] In recent years, the correlation between infrastructure issues and events and weather information has increased due to emergency measures in response to the normalization of disasters caused by abnormal weather, and the introduction of renewable energy (renewable energy generators) as a management strategy for power companies and industrial consumers. For example, advanced technology is required for predicting the amount of power generated by photovoltaic (PV) power generation facilities for renewable energy. However, low accuracy in power generation forecasts can lead to imbalances not being met in power aggregation businesses and poor profits due to inappropriate bidding in power trading. Therefore, by implementing highly accurate PV power generation forecasts, it will ultimately be possible to reduce imbalance costs in renewable energy balancing and contribute to appropriate power trading.

[0003] PV power generation forecasts are often carried out using, for example, image data captured by meteorological satellites and solar radiation data as inputs, without taking into account the state of the PV panels. However, when snow accumulates on PV panels, the amount of solar radiation is blocked, resulting in a decrease in PV power generation. Therefore, if snow accumulation on PV panels is not taken into account, there is a possibility that a large discrepancy will occur between the predicted power generation value and the actual value. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 6852621 Summary of the Invention [Problem to be solved by the invention]

[0005] As a countermeasure, for example, there is technology that predicts the snow accumulation on PV panels using information mainly on snow depth and the number of days that have passed since snowfall ended.

[0006] In addition, observing the depth of snow on PV panels requires equipment and personnel costs. Therefore, it is possible to predict snow accumulation on PV panels using photographic image data and weather forecast data, but if these data contain missing values ​​or if there is a time gap between the base point and the target point in time, the prediction accuracy may be poor.

[0007] Therefore, the embodiments of the present invention provide a photovoltaic power generation management device, a photovoltaic power generation management method, and a photovoltaic power generation management program that are capable of predicting the state of snow accumulation on a PV panel with high accuracy. [Means for solving the problem]

[0008] According to one embodiment, the photovoltaic power generation management device includes an acquisition unit that acquires photographed image data of a predetermined area including an installation location of a PV panel at a reference time point and weather forecast data at a prediction target time point. The photovoltaic power generation management device further includes a snowfall prediction unit that predicts the snowfall condition of the predetermined area at the prediction target time point based on the photographed image data and weather forecast data acquired by the acquisition unit. The photovoltaic power generation management device further includes a snowfall prediction unit that predicts the snowfall condition on the PV panel at the prediction target time point based on the prediction result of the snowfall condition. The acquisition unit further checks whether there is a missing part in the weather forecast data and acquires weather forecast data for a timing prior to the prediction target time point. [Brief description of the drawings]

[0009] [Figure 1] 1 is a schematic configuration diagram of a solar power generation management system 10 according to a first embodiment. [Diagram 2] 1 is a block diagram of a solar power generation management device 1 according to a first embodiment. [Diagram 3] 1 is a flow chart showing a process in which the photovoltaic power generation management device 1 in the first embodiment predicts a power generation amount while taking into account missing data. [Figure 4]4 is an example in which the power generation amount prediction results in the first embodiment are output in chronological order. [Diagram 5] 13 is a flow diagram showing a process in which the photovoltaic power generation management device 1 in the first embodiment performs snowfall prediction in consideration of snowfall prediction data before a time point of prediction. [Figure 6] 13 is a flow chart for collecting correct answer data of a snow accretion learning model based on an actual power generation amount value and a predicted power generation amount value in the second embodiment. [Figure 7] 10 is a flowchart showing a process performed by a snow accretion prediction unit 23 in the second embodiment when classifying the presence or absence of snow accretion on a PV panel 70. [Figure 8] 13 is an example in which the state of snow accumulation on the PV panel 70 and the actual power generation amount value and the like are output in chronological order in the second embodiment. [Figure 9] FIG. 11 is a block diagram of a photovoltaic power generation management device 1 according to a third embodiment. [Figure 10] 13 is a flow chart showing a process in which the photovoltaic power generation management device 1 according to the third embodiment predicts snow accretion by threshold value judgment. [Figure 11] FIG. 13 is a hardware configuration diagram of a photovoltaic power generation management device 1 in a fourth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Hereinafter, an embodiment of the present disclosure will be described with reference to the drawings. The present embodiment does not limit the present invention. The drawings are schematic or conceptual, and the ratio of each part is not necessarily the same as the actual one. In the specification and drawings, elements similar to those described above with respect to the previous drawings are given the same reference numerals, and detailed descriptions are omitted as appropriate.

[0011] In this embodiment, the state in which snow adheres to the PV panels is referred to as "snow accumulation," and the state in which snow accumulates on other surfaces such as the ground is referred to as "snow accumulation."

[0012] (First embodiment) FIG. 1 is a schematic configuration diagram of a photovoltaic power generation management system 10 according to the first embodiment.

[0013] The photovoltaic power generation management system 10 includes a photovoltaic power generation management device 1, a satellite image server 40, a weather server 41, a network device 50, a weather satellite 60, a receiving antenna 61, and a PV panel 70.

[0014] 1, a string configuration 80 of PV panels 70 includes multiple PV panels 70 connected in series with each other, and an array configuration 90 of PV panels 70 includes multiple string configurations 80 connected in parallel with each other.

[0015] The photovoltaic power generation management device 1 in this embodiment is installed, for example, in a management center 100 that remotely manages the photovoltaic power plant 200, and performs predictions of snow accumulation in a predetermined area including the installation location of the PV panels 70, predictions of snow accumulation on the PV panels 70, and predictions of power generation amount taking into account snow accumulation on the PV panels 70. In addition to this example, the photovoltaic power generation management device 1 may be installed in a building or the like within the premises of the photovoltaic power plant 200.

[0016] Further, in this example, the power generation amount prediction is performed for one photovoltaic power plant 200, but the photovoltaic power generation management device 1 may be configured to perform power generation amount prediction for a plurality of photovoltaic power plants 200.

[0017] Hereinafter, the installation location of the PV panel 70 is also simply referred to as the "installation location", and the predetermined area including the installation location of the PV panel 70 for which the snow accumulation condition is predicted is also simply referred to as the "predetermined area".

[0018] The photovoltaic power generation management device 1 in this embodiment also acquires photographed image data of a predetermined area photographed by a meteorological satellite 60 from the satellite image server 40 via the network 51. The photovoltaic power generation management device 1 also acquires weather forecast data (temperature data, humidity data, solar radiation data, etc.) of the predetermined area from the weather server 41 via the network 51. The photographed image data and the weather forecast data are periodically transmitted from the satellite image server 40 and the weather server 41.

[0019] Moreover, the photovoltaic power generation management device 1 in this embodiment acquires weather record data (temperature data, humidity data, solar radiation amount data, etc.) from the weather server 41. The weather record data is transmitted from the weather server 41 periodically.

[0020] Image data of a predetermined area is captured by a meteorological satellite 60. The captured data is received by a receiving antenna 61 via a satellite network 62 and stored in the satellite image server 40.

[0021] The captured image data and weather forecast data may be data provided by, for example, a government agency or a private company. In such a configuration, the satellite image server 40 and the weather server 41 are constructed by the government agency, private company, or the like that provides the data. The photovoltaic power generation management device 1 may acquire the captured image data and weather forecast data via the network 51, for example, by using a Web API (Application Programming Interface).

[0022] In addition, the network 51 described in this embodiment will be described using an example of an Internet line connected via the network device 50, but for example, some or all of the sections of the network 51 connecting the management center 100 and the solar power plant 200 may be constructed using a LAN (Local Area Network).

[0023] FIG. 2 is a block diagram of the photovoltaic power generation management device 1 in the first embodiment.

[0024] In the following, the reference time point is also referred to as the reference time point, and the time point after the reference time point for which the snow accumulation condition is to be predicted is also referred to as the prediction target time point. For example, the reference time point may be the prediction execution time point, or may be the shooting time point of the photographed image data used to estimate the snow accumulation condition.

[0025] The photovoltaic power generation management device 1 in this embodiment includes a processing unit 2, a storage unit 3, an input unit 4, an output unit 5 and a communication unit 6.

[0026] Note that some of the functions of the photovoltaic power generation management device 1 can be realized by, for example, cloud computing. However, in this embodiment, for the sake of simplicity, the photovoltaic power generation management device 1 will be described as being a single computer device.

[0027] The communication unit 6 is a communication interface for communicating (transmitting and receiving) various information with external devices, various sensors in the power system, etc. The photovoltaic power generation management device 1 acquires the above-mentioned captured image data, weather forecast data, and weather record data via the communication unit 6.

[0028] The storage unit 3 is realized by, for example, a flash memory, a hard disk drive (HDD), a solid state drive (SSD), etc. The storage unit 3 stores, for example, the operation program of the processing unit 2, various data, various calculation results by the processing unit 2, etc. The storage unit 3 may be realized by a random access memory (RAM) or a read only memory (ROM).

[0029] Specifically, the storage unit 3 stores, for example, as data to be used for predicting the state of snow accretion on PV panels, the photographed image data and weather forecast data acquired by the communication unit 6, data for threshold determination, data for machine learning, equipment information data (latitude and longitude information of the installation point, tilt angle information of the PV panel, panel azimuth angle information, etc.), and time characteristic information regarding the prediction target time point (information indicating characteristics (temperature, humidity, amount of solar radiation, etc.) of the season, time period (morning, noon, night, etc.)), etc. The data stored in the storage unit 3 is stored together with time information data indicating the time of the data.

[0030] The storage unit 3 also stores captured image data in a weather satellite database, weather forecast data in a weather forecast database, and facility information data in a facility information database, which will be described later. The storage unit 3 also stores, in a PV power generation prediction database, predicted power generation values ​​in a case where snow accretion on the PV panel 70 is taken into consideration and a case where snow accretion on the PV panel 70 is not taken into consideration, respectively. The storage unit 3 also stores, in a past performance database, actual power generation values ​​of the PV panel 70, actual snow accretion data on the PV panel 70, and actual weather data, as well as facility information data at the time points corresponding to each actual data.

[0031] The processing unit 2 is realized, for example, by a CPU (Central Processing Unit). The processing unit 2 operates based on the operation program and various data stored in the storage unit 3, thereby realizing, as functional units, an acquisition unit 21, a snow accumulation prediction unit 22, a snow accumulation prediction unit 23, and a power generation prediction unit 24. More specifically, the CPU deploys the operation program of the processing unit 2 stored in a HDD or the like on a RAM and executes it, thereby realizing the functions of each functional unit.

[0032] The acquisition unit 21 acquires various information from the storage unit 3, an external device, etc. The acquisition unit 21 acquires captured image data, weather forecast data, facility information data, etc. from the storage unit 3 as data used for snowfall prediction, snow accumulation prediction, etc.

[0033] Due to malfunctions or maintenance of meteorological satellites or sensors, photographed image data or weather forecast data may be temporarily missing. In addition, there may be cases where the files storing photographed image data or weather forecast data do not exist, and the data does not exist for a certain period of time. In the following, for the sake of explanation, the case where data such as photographed image data or weather forecast data is temporarily missing or where the files storing these data do not exist is referred to as data being missing.

[0034] In this embodiment, the acquisition unit 21 checks whether there is any missing data when acquiring photographed image data, weather forecast data, equipment information data, etc. used in snowfall prediction and snow accumulation prediction. If photographed image data at the reference time point is missing, the acquisition unit 21 acquires photographed image data captured at a timing earlier than the reference time point. Furthermore, if photographed image data at the prediction target time point is missing, the acquisition unit 21 acquires weather forecast data at a timing earlier than the prediction target time point. Furthermore, if equipment information data at the prediction target time point is missing, the acquisition unit 21 acquires equipment information data at a timing earlier than the prediction target time point.

[0035] For example, if captured image data or weather forecast data is missing, the acquisition unit 21 acquires and substitutes the most recent captured image data or weather forecast data in place of the missing data. If weather forecast data is transmitted from the weather server 41 once an hour and weather forecast data for 11:00 a.m. is missing, the acquisition unit 21 acquires weather forecast data for 10:00 a.m., which is the most recent time in the past. If data is continuously missing, the acquisition unit 21 acquires weather forecast data going back to the time when data exists. The acquisition unit 21 also acquires time information data of the weather forecast data.

[0036] Furthermore, the acquisition unit 21 may use captured image data captured at a past time point that is similar to the weather conditions at the reference time point, for example, as a substitute for the missing captured image data.

[0037] The snowfall prediction unit 22 predicts the snowfall conditions in a specific area at a prediction target time based on photographed image data of the specific area and weather prediction data, and outputs a prediction result of the snowfall conditions (also simply called a snowfall prediction result).

[0038] For example, the snowfall prediction unit 22 estimates the snowfall conditions in a specified area at a reference time based on the captured image data, and predicts the snowfall conditions in the specified area at the prediction target time using threshold judgment or machine learning based on the estimated snowfall conditions (also simply called the snowfall estimation result) and weather forecast data, and outputs information indicating the snow depth or the presence or absence of snowfall as a snowfall prediction result.

[0039] Here, we will explain the estimation of snow conditions in a given area at a reference time based on captured image data. By subjecting the captured image data to image synthesis processing (e.g., RGB (Red, Green, Blue) synthesis processing), it becomes possible to determine the presence or absence of snow on the ground, as well as vegetation and clouds in the sky. For example, snow has a high reflectance to visible light and a high absorptance to near-infrared light, while clouds have a high reflectance to both visible light and near-infrared light, and these characteristics allow snow and clouds to be distinguished.

[0040] The image synthesis processing technique (recipe) used may be, for example, one that results in a synthesis of visible light or one that is used to distinguish between snow and fog, but it is also possible to use a combination of multiple techniques.

[0041] Next, we will explain the prediction of snow accumulation conditions in a specific area at a target time based on the snow accumulation estimation result in the specific area at the reference time and the weather forecast data at the target time. For this prediction, for example, two types of methods, threshold judgment and machine learning, can be used.

[0042] In the case of threshold judgment, a parameter is used to judge the threshold. In the case of machine learning, for example, random forest, regression analysis, logistic regression, etc. are used. In addition, snowfall prediction results can be, for example, two types of information indicating snow depth or the presence or absence of snow, but it is also possible to finally set a threshold value for snow depth and make it a binary classification of the presence or absence of snow.

[0043] In the machine learning, the snowfall prediction unit 22 uses a snowfall prediction model. The snowfall prediction model receives as input a snowfall estimation result at a reference time point and weather forecast data at a prediction target time point, and outputs information indicating snowfall depth or the presence or absence of snowfall.

[0044] In the threshold judgment, the snow accumulation prediction unit 22 judges whether or not there is snow accumulation from the snow accumulation estimation result, and if it judges that there is snow accumulation, it compares the weather forecast data, which is a parameter, with a threshold value and outputs information indicating the snow depth or the presence or absence of snow accumulation.

[0045] In addition, when comparing threshold judgment and machine learning, it is considered that machine learning has higher accuracy when the amount of learning data for the snowfall prediction model in machine learning is relatively large, and threshold judgment has higher accuracy when the amount of learning data is relatively small. This is because if the amount of learning data is insufficient, the snowfall prediction model may become over-learned. Therefore, the amount of learning data at that boundary is set as the first predetermined value.

[0046] The snowfall prediction unit 22 uses threshold judgment, for example, if the amount of learning data of the snowfall prediction model in machine learning is less than a first predetermined value, and uses machine learning, for example, if the amount of learning data of the snowfall prediction model in machine learning is equal to or greater than the first predetermined value. Specifically, for example, for a point where data is difficult to obtain, threshold judgment is used if the amount of learning data is less than the first predetermined value. Also, it is possible to perform snowfall prediction using threshold judgment when the amount of learning data is less than the first predetermined value immediately after the system is introduced, and switch the prediction method to machine learning after a predetermined period of time has passed and the amount of learning data becomes equal to or greater than the first predetermined value. In this way, threshold judgment and machine learning can be appropriately used depending on the amount of learning data.

[0047] The scope of the snowfall prediction model can be determined arbitrarily. That is, for example, it may be a uniform model for the whole country, or it may be a model for each area. Here, the area may be a division of administrative units such as a city, ward, town, village, prefecture, or region, or it may be arbitrarily set by the developer.

[0048] For example, a business operator who has installed multiple solar power plants in a prefecture can create a snowfall forecast model for all the power plants in the prefecture at once. The business operator can also designate area divisions within the prefecture and create snowfall forecast models for each area. However, when arbitrarily setting areas, the condition is that weather forecast data for at least one location in the area can be prepared.

[0049] In addition, since the snowfall prediction unit 22 predicts the snowfall conditions in a specified area at the prediction target time based on the snowfall estimation results for the specified area at the reference time and weather forecast data as described above, if there is a time gap between the reference time and the prediction target time, the prediction accuracy decreases.

[0050] Therefore, in this embodiment, if the reference time is t=0 and the prediction time is t=s, the snowfall prediction unit 22 predicts snowfall at the prediction time by further using the snowfall prediction results at times t=0, sn, ..., s-1 (n is any natural number), that is, at the timing between the reference time and the prediction time (before the prediction time).

[0051] When estimating snowfall at time t=sn, the snowfall prediction unit 22 uses the snowfall estimation result at time t=0 and the weather forecast data at time t=sn. Similarly, when estimating snowfall at time t=s-n+1, ​​the snowfall prediction unit 22 uses the snowfall estimation result at time t=0, the snowfall forecast result at time t=sn, and the weather forecast data at time t=s-n+1. Through these procedures, the snowfall prediction unit 22 calculates the snowfall forecast result at time t=s-1.

[0052] The snow accumulation prediction unit 23 predicts the snow accumulation state on the PV panel 70 at the prediction target time based on the snow accumulation prediction result at the time t=s, and outputs the snow accumulation prediction result.

[0053] For example, the snow accumulation prediction unit 23 predicts the snow accumulation condition on the solar power generation panel at the predicted time point using threshold judgment or machine learning based on at least one of the weather forecast data, equipment information data, and time characteristic information related to the predicted time point, in addition to the snow accumulation prediction result, and outputs information indicating the presence or absence of snow accumulation or information indicating the probability of snow accumulation as the snow accumulation prediction result.

[0054] In the machine learning, the snowfall prediction unit 23 uses a snowfall prediction model. The snowfall prediction model receives, for example, at least one of a snowfall prediction result at a prediction target time, weather prediction data, facility information data, and time characteristic information related to a prediction target time as input, and outputs information indicating the presence or absence of snowfall or information indicating the probability of snowfall.

[0055] In the threshold judgment, the snow accumulation prediction unit 23, for example, judges whether or not there is snow accumulation from the snow accumulation forecast result, and if it judges that there is snow accumulation, compares at least one of the parameters, namely, the weather forecast data, the equipment information data, and the time characteristic information related to the prediction target time point, with a threshold value, and outputs information indicating the presence or absence of snow accumulation or information indicating the probability of snow accumulation.

[0056] Furthermore, the snowfall prediction unit 23 uses threshold judgment if the amount of learning data of the snowfall learning model in machine learning is less than a second predetermined value, and uses machine learning if the amount of learning data of the snowfall learning model in machine learning is equal to or greater than the second predetermined value. As in the case of snowfall prediction, if the amount of learning data is insufficient, the model may become over-learned, so the amount of learning data at the boundary is set to the second predetermined value, and threshold judgment and machine learning are appropriately used. The same method as that of the snowfall prediction unit 22 can be used to distinguish between threshold judgment and machine learning.

[0057] The power generation prediction unit 24 predicts the power generation amount at the prediction target time point using weather forecast data, photographed image data, snow accretion prediction results, equipment information data, etc. For example, the power generation prediction unit 24 predicts the power generation amount using solar radiation data as weather forecast data, tilt angle information of the PV panel 70 as equipment information data, and further using snow accretion prediction results. For example, the power generation prediction unit 24 predicts the power generation amount using information indicating the presence or absence of snow accretion or information indicating the probability of snow accretion in addition to the solar radiation data at the prediction target time point and the tilt angle of the solar panel acquired as equipment information data. The result of the power generation prediction is output as a power generation prediction value together with time information data at the prediction target time point.

[0058] In addition, as will be described in detail later, when outputting the power generation prediction result, the power generation prediction unit 24 judges whether the prediction result has been created using weather forecast data or the like without any missing data, and changes the output method of the prediction result depending on the judgment result. In this example, if the power generation prediction result has been created using weather forecast data or the like without any missing data, the power generation prediction unit 24 outputs the power generation prediction result as time series data indicated by a solid line, and if the prediction result has been created using weather forecast data or the like with missing data, the power generation prediction unit 24 outputs the power generation prediction result as time series data indicated by a broken line. For example, the power generation prediction unit 24 judges whether the power generation prediction result has been created using missing data from the time information data of the data used for the power generation prediction.

[0059] The prediction of the amount of power generation may be performed for each PV panel 70 for which the presence or absence of snow has been determined, or may be performed for each string configuration 80 or array configuration 90, or for the entire solar power plant 200.

[0060] The input unit 4 is a means for a user to input information, such as a mouse, a keyboard, or a touch panel.

[0061] The output unit 5 is a display device such as an LCD (Liquid Crystal Display) or an audio output device such as a speaker. For example, the output unit 5 outputs the power generation amount prediction result by the power generation amount prediction unit 24.

[0062] FIG. 3 is a flow chart showing how the photovoltaic power generation management device 1 in the first embodiment predicts the amount of power generation while taking into account missing data.

[0063] In the following, for the sake of simplicity, the flow for performing power generation prediction taking into account missing data and the flow for performing snowfall prediction taking into account snowfall prediction data prior to the prediction time will be explained separately.

[0064] In this example, a facility information database 31, a weather satellite database 32, and a weather forecast database 33 are illustrated among the databases included in the storage unit 3. The facility information data, captured image data, and weather forecast data will be described as being stored in advance in the databases. For the sake of explanation, in the flow of FIG. 3, the snowfall prediction unit 22 that estimates the snowfall condition in a specified area at a reference time point is referred to as a snowfall prediction unit 22', and the snowfall prediction unit 22 that predicts snowfall at a prediction target time point is referred to as a snowfall prediction unit 22''.

[0065] The acquisition unit 21 acquires facility information data from the facility information database 31, acquires meteorological satellite data from the meteorological satellite database 32, and acquires weather forecast data from the weather forecast database 33 as data used in snowfall forecast and snowfall forecast. The acquisition unit 21 checks whether there is any missing data when acquiring these data. The lower left of FIG. 3 shows an example in which the acquisition unit 21 confirms that humidity data and solar radiation data at 1:00 on January 2, 2022 are missing from the acquired weather forecast data. The acquisition unit 21 acquires the most recent data in the past instead of the data of the missing portion. In the example of FIG. 3, the acquisition unit 21 acquires humidity data and solar radiation data at 0:00 on January 2, 2022 as the most recent humidity data and solar radiation data in the past instead of the humidity data and solar radiation data of the missing portion.

[0066] In addition, when acquiring the weather forecast data and meteorological satellite data, the acquiring unit 21 also acquires time information data for the data. By acquiring the time information data, the snow accumulation predictor 22, the snow accumulation predictor 23, and the power generation predictor 24 can determine whether the data has been substituted with past data.

[0067] The snowfall prediction unit 22' receives input of weather forecast data and meteorological satellite data acquired by the acquisition unit 21. At a reference time point, the snowfall prediction unit 22 estimates the snowfall conditions in a specified area using the photographed image data at the reference time point acquired by the acquisition unit 21. Furthermore, the snowfall prediction unit 22'' predicts the snowfall conditions in a specified area at a prediction target time point using the snowfall estimation result at the reference time point and the weather forecast data at the prediction target time point, and outputs the snowfall prediction result.

[0068] The snow accumulation prediction unit 23 receives the snow accumulation prediction result, weather prediction data, facility information data, etc. as input, and outputs the snow accumulation prediction result at the prediction target time.

[0069] The power generation prediction unit 24 outputs the power generation prediction result at the prediction target time point using weather forecast data, photographed image data, snowfall prediction results, facility information data, etc. In this example, the power generation prediction unit 24 determines that the power generation prediction was performed using past data from the time information data of the weather forecast data out of the respective time information data of the weather forecast data, the photographed image data, and the facility information data. Also, in this example, the power generation prediction unit 24 outputs the power generation prediction value together with the time information data of the prediction target time.

[0070] FIG. 4 is an example in which the power generation amount prediction results in the first embodiment are output in chronological order.

[0071] 4, the horizontal axis indicates the prediction target time point (shown in date and time), and the vertical axis indicates the predicted power generation amount (shown in kW). In this embodiment, even if there is a loss in the weather prediction data and the captured image data, the snow accumulation prediction unit 22, the snow accumulation prediction unit 23, and the power generation amount prediction unit 24 supplement the weather prediction data and the captured image data using the most recent past data, so that it is possible to predict the power generation amount even if there is a loss.

[0072] The power generation amount prediction unit 24 judges whether the predicted power generation amount is created by substituting past data for the data of the missing portion from the time information data. In the example of FIG. 4, the result predicted using the weather forecast data and the photographed image data without missing data is shown by a solid line, and the result predicted by substituting past data due to missing data is shown by a dashed line. The predicted power generation amount is shown by a solid line from January 27th 0:00 to January 28th 0:00 and from January 29th 0:00 to January 30th 0:00, and the predicted power generation amount is shown by a dashed line from January 28th 0:00 to January 29th 0:00. The user can check whether the predicted power generation amount is a value predicted by substituting past data by checking the type of line of the output predicted power generation amount value.

[0073] FIG. 5 is a flow chart showing a process in which the photovoltaic power generation management device 1 in the first embodiment performs snowfall prediction in consideration of snowfall prediction data before the prediction target time point.

[0074] In this example, a facility information database 31, a weather satellite database 32, and a weather forecast database 33 are illustrated among the databases included in the storage unit 3. The facility information data, captured image data, and weather forecast data will be described as being stored in advance in the respective databases. Also, in this example, the description will be made on the assumption that there is no loss of any of the data.

[0075] As described above, in order to prevent a decrease in prediction accuracy due to the time gap between the reference time point and the prediction target time point, the snowfall prediction unit 22 further uses the snowfall prediction results at times t = 0, sn, ..., s-1 (n is any natural number) to predict snowfall at the prediction target time point.

[0076] 5, for the sake of explanation, a flow for predicting snowfall at time t=s will be described using snowfall prediction results at three time points, t=0, t=sn, and t=s-1. Also, the snowfall prediction unit 22 at time t=0 will be referred to as snowfall prediction unit 22a, the snowfall prediction unit 22 at time t=sn will be referred to as snowfall prediction unit 22b, the snowfall prediction unit 22 at time t=s-1 will be referred to as snowfall prediction unit 22c, and the snowfall prediction unit 22 at time t=s will be referred to as snowfall prediction unit 22d.

[0077] The snowfall prediction unit 22a receives an input of meteorological satellite data acquired by the acquisition unit 21. At a reference time point, the snowfall prediction unit 22a estimates the snowfall condition in a predetermined area by using the photographed image data at the reference time point acquired by the acquisition unit 21.

[0078] In addition, the snowfall prediction unit 22b predicts the snowfall conditions at t=sn before the prediction target time using the snowfall estimation result at the reference time t=0 and weather prediction data at t=sn before the prediction target time, and outputs the snowfall prediction result.

[0079] In addition, the snowfall prediction unit 22c predicts the snowfall conditions at t=s-1 before the prediction time point using the snowfall estimation result at the reference time point t=0, the snowfall prediction result at t=sn before the prediction time point, and weather prediction data at t=s-n+1 before the prediction time point, and outputs the snowfall prediction result.

[0080] In addition, the snowfall prediction unit 22d predicts the snowfall conditions at the prediction time t=s using the snowfall estimation result at the reference time t=0, the snowfall prediction result at t=s before the prediction time t=s, the snowfall prediction result at t=s-1 before the prediction time t=s, and the weather prediction data for the prediction time t=s, and outputs the snowfall prediction result.

[0081] The snow accumulation prediction unit 23 predicts the snow accumulation state of the PV panel 70 based on the snow accumulation prediction result at the prediction target time t=s, as described above, and outputs the snow accumulation prediction result. The power generation amount prediction unit 24 outputs the power generation amount prediction result at the prediction target time, as described above, using the weather prediction data, the captured image data, the snow accumulation prediction result, the facility information data, etc.

[0082] According to this embodiment, even if there is a loss in the weather forecast data, the photovoltaic power generation management device 1 substitutes past data for snowfall prediction and performs snowfall prediction. As a result, even if there is a loss in the weather forecast data, the photovoltaic power generation management device 1 can perform highly accurate power generation prediction taking into account the snow accumulation on the PV panel 70.

[0083] According to the present embodiment, the photovoltaic power generation management device 1 performs snowfall prediction using the snowfall prediction result before the prediction target time point. This enables the photovoltaic power generation management device 1 to perform snowfall prediction with high accuracy even if there is a large time gap between the reference time point and the prediction target time point.

[0084] Second embodiment FIG. 6 shows a flow for collecting correct answer data of a snow accretion learning model based on the actual power generation amount value and the predicted power generation amount value in the second embodiment.

[0085] In this embodiment, the system configuration diagram of the photovoltaic power generation management system 10 is the same as that in FIG. 1, and the block diagram of the photovoltaic power generation management device 1 is the same as that in FIG. 2, so that the description will be omitted.

[0086] When creating a snow accretion learning model for each photovoltaic power plant 200, if a camera is installed on-site or someone is dispatched to the site to measure the presence or absence of snow accretion on the PV panels 70, a cost is incurred to obtain correct data. The cost is particularly high when creating snow accretion learning models for multiple photovoltaic power plants 200. Therefore, in this embodiment, the snow accretion prediction unit 23 creates a snow accretion learning model based on the actual power generation amount value of the PV panel 70 and the power generation amount prediction value when snow accretion on the PV panel 70 is not taken into consideration.

[0087] In this example, of the databases included in the storage unit 3, a past performance database 34 and a PV power generation amount prediction database 35 are illustrated.

[0088] As described above, the past performance database 34 stores the actual power generation amount value of the PV panel 70. This actual value may be the actual power generation amount value for each PV panel 70, or may be the actual power generation amount value for the entire photovoltaic power plant 200. In addition, this actual power generation amount value is acquired from the PV panel 70 by the photovoltaic power generation management device 1 via the network 51, for example. In addition, the actual power generation amount value is stored together with the corresponding time information data.

[0089] As described above, the PV power generation amount prediction database 35 stores the power generation amount predicted value when the power generation amount prediction unit 24 does not take into account the accumulation of snow on the PV panel 70. The power generation amount predicted value when the accumulation of snow on the PV panel 70 is not taken into account is generated, for example, by the power generation amount prediction unit 24. Furthermore, the power generation amount predicted value when the accumulation of snow on the PV panel 70 is not taken into account is stored together with the corresponding time information data.

[0090] The snow accretion prediction unit 23 acquires the power generation amount actual value from the past performance database 34, and acquires the power generation amount predicted value in the case where snow accretion on the PV panel 70 is not taken into consideration from the PV power generation amount prediction database 35. At this time, the snow accretion prediction unit 23 acquires the power generation amount actual value that matches the time information data and the power generation amount predicted value in the case where snow accretion on the PV panel 70 is not taken into consideration.

[0091] The snow accretion classification unit 231 in the snow accretion prediction unit 23 classifies the presence or absence of snow accretion on the PV panel 70 using the acquired power generation prediction value and power generation actual value when snow accretion on the PV panel 70 is not taken into consideration. For example, the snow accretion prediction unit 23 further acquires snow accumulation actual data from the past actual data database 34. When the snow accretion prediction unit 23 determines that the reduction rate of the power generation actual value with respect to the power generation prediction value when snow accretion on the PV panel 70 is not taken into consideration is equal to or greater than a predetermined value at a date and time when the snow depth is 1 cm or more, it can be said that power generation is hindered by snow accumulation on the PV panel 70, so it classifies that there is snow accretion on the PV panel 70, and when it is less than the predetermined value, it classifies that there is no snow accretion on the PV panel. In this way, the snow accretion learning model can use data classified as to the presence or absence of snow accretion on the PV panel 70 (snow accretion classification data) as the correct answer data of the snow accretion learning model.

[0092] Furthermore, for example, the snow accumulation prediction unit 23 generates snow accumulation presence / absence classification data by assigning a flag corresponding to the actual data regarding the presence or absence of snow accumulation on the PV panel 70. For example, a snow accumulation flag of 1 is assigned to the actual snow accumulation data when the actual snow accumulation is classified as 1 cm or more, and 0 otherwise. Furthermore, the snow accumulation prediction unit 23 assigns a snow accumulation influence flag of 1 when it is classified as having snow accumulation on the PV panel 70, and 0 otherwise, to a pair of the actual power generation amount value and the predicted power generation amount value without considering snow accumulation on the PV panel 70. The snow accumulation prediction unit 23 stores the data with these flags in the past performance database 34 as snow accumulation presence / absence classification data.

[0093] FIG. 7 is a flowchart showing a process performed by the snow accumulation prediction unit 23 in the second embodiment to classify the presence or absence of snow accumulation on the PV panel 70.

[0094] In step S1, the acquisition unit 21 acquires a power generation amount actual value from the past performance database 34 in the storage unit 3, and acquires a power generation amount predicted value without considering snow accumulation on the PV panel 70 from the PV power generation amount prediction database 35. The values ​​acquired at this time are the values ​​of the date and time for which the snow accumulation prediction unit 23 judges whether or not snow has accumulated on the PV panel 70. The acquisition unit 21 also acquires past snow accumulation data from the past performance database 34. The past snow accumulation data acquired at this time are the values ​​of the date and time for which the snow accumulation prediction unit 23 judges whether or not snow has accumulated on the PV panel 70. In step S2, the snow accumulation prediction unit 23 determines whether or not the snow accumulation data is N cm or more (N is a positive value). It is considered that N used at this time is input in advance by the user using the input unit 4 via a user interface displayed on the output unit 5, for example. The user interface for input may be provided by a setting screen.

[0095] If the snow accumulation data is equal to or greater than N (yes in step S2), in step S3, a snow accumulation flag value of 1 is assigned to the acquired power generation amount actual value. In step S4, the snow accumulation prediction unit 23 judges whether or not the rate of decrease in the power generation amount actual value is equal to or greater than M% with respect to the power generation amount predicted value when snow accumulation on the PV panel 70 is not taken into consideration. The judgment formula at this time is expressed by formula (1). Similarly to N, M used at this time can be input in advance by the user using the input unit 4 via a user interface displayed on the output unit 5, for example. Note that in formula (1), the power generation amount predicted value when snow accumulation on the PV panel 70 is not taken into consideration is simply expressed as the power generation amount predicted value. (Predicted power generation – Actual power generation) / Predicted power generation >= M (1)

[0096] If the decrease rate of the actual power generation value compared to the predicted power generation value when snow accretion on the PV panel 70 is not taken into account is M% or more (yes in step S4), in step S5, the snow accretion prediction unit 23 assigns a snow accretion impact flag of 1 to the pair of the actual power generation value and the predicted power generation value when snow accretion on the PV panel 70 is not taken into account, and terminates the processing.

[0097] On the other hand, if the actual snow accumulation is not N cm or more (no in step S2), in step S6, the snow accretion prediction unit 23 assigns a snow accumulation flag value of 0 to the actual power generation amount value. After that, in step S7, the snow accretion prediction unit 23 assigns a snow accretion influence flag of 0 to the pair of the actual power generation amount value and the power generation amount predicted value in the case where snow accretion on the PV panel 70 is not taken into consideration, and ends the process.

[0098] If the decrease rate of the actual power generation value relative to the predicted power generation value when snow accretion on the PV panel 70 is not M% or more (no in step S4), in step S7, the snow accretion prediction unit 23 assigns a snow accretion impact flag of 0 to the pair of the actual power generation value and the predicted power generation value when snow accretion on the PV panel 70 is not taken into account, and ends the processing.

[0099] FIG. 8 is an example of a chronological output of the state of snow accumulation on the PV panel 70 and actual power generation amount values ​​in the second embodiment.

[0100] 8, the background of the period when the snow accumulation flag was 1 is shown in light gray, and the background of the period when the snow accumulation impact flag was 1 is shown in dark gray. The actual power generation amount is shown in a solid line, and the predicted power generation amount when snow accumulation on the PV panel 70 is not taken into consideration is shown in a dashed line. As shown in FIG. 8, when there is a large difference at a certain point in time between the actual power generation amount and the predicted power generation amount when snow accumulation on the PV panel 70 is not taken into consideration, this indicates that the rate of decrease in the predicted power generation amount is large compared to the actual power generation amount, and it can be seen that the snow accumulation impact flag is set to 1.

[0101] According to this embodiment, the photovoltaic power generation management device 1 classifies the presence or absence of snow accumulation on the PV panel 70 at any time based on the actual snow accumulation data, the actual power generation amount value, and the predicted power generation amount value without considering snow accumulation on the PV panel 70. This makes it possible to obtain correct data on the presence or absence of snow accumulation on the PV panel 70 without incurring costs for installing equipment such as cameras or dispatching personnel, and makes it easy to create a snow accumulation learning model for each photovoltaic power plant 200.

[0102] Third embodiment FIG. 9 is a block diagram of a photovoltaic power generation management device 1 in the third embodiment.

[0103] In this embodiment, an example will be described in which the photovoltaic power generation management device 1 uses threshold judgment out of two methods, threshold judgment and machine learning, to predict snow accretion on the PV panel 70. When predicting snow accretion, as described above, the photovoltaic power generation management device 1 uses weather forecast data, facility information data, and the like as parameters and sets thresholds for these parameters to predict snow accretion.

[0104] Similarly to the first embodiment, the photovoltaic power generation management device 1 in this embodiment includes a processing unit 2, a storage unit 3, an input unit 4, an output unit 5, and a communication unit 6. Hereinafter, a description of the same matters as in the first embodiment will be omitted.

[0105] Unlike the first embodiment, the photovoltaic power generation management device 1 in this embodiment includes, as functional units, an acquisition unit 21, a snow accumulation prediction unit 22, a snow accumulation prediction unit 23, and a power generation amount prediction unit 24, as well as a threshold determination unit 25. The CPU deploys on the RAM an operation program of the processing unit 2 stored in the HDD or the like and executes it, thereby realizing the function of each functional unit.

[0106] The threshold determination unit 25 determines the threshold of each parameter when performing threshold determination in snow accretion prediction. In this example, an example of determining thresholds related to the weather forecast data and facility information data described above will be described, but the threshold determination unit 25 may be configured to determine various other thresholds for determining snow accretion on the PV panel 70.

[0107] The threshold determination unit 25 may input, for example, snow accumulation record data on the PV panel 70, weather record data, snow accumulation record data, and equipment information data at the time points corresponding to these record data, and determine each threshold using machine learning, and output it as threshold data. The snow accumulation record data on the PV panel 70 used in the machine learning may use the snow accumulation presence / absence classification data described in the second embodiment. The threshold determination model in the machine learning inputs, for example, snow accumulation record data on the PV panel 70, weather record data, snow accumulation record data, and equipment information data at the time points corresponding to these record data, and outputs threshold data for predicting the snow accumulation state of the weather forecast data or equipment information data, etc., which are parameters. The threshold determination unit 25 uses, for example, equipment information data whose time information data matches that of the record data used in the threshold determination, as equipment information data at the time points corresponding to each record data.

[0108] In addition, an example of correcting the thresholds of weather forecast data using inclination angle information of the PV panel 70, which is equipment information data, will be described below. Here, an example of the temperature threshold T and solar radiation threshold R among the thresholds of the weather forecast data will be described. There is a tendency that snow that has adhered to the PV panel 70 is more likely to fall off if the inclination angle of the panel is high. In order to reflect this tendency in the thresholds of the weather forecast data, the threshold determination unit 25 determines the temperature threshold T' and solar radiation threshold R', which are corrected thresholds, for example, by using equations (2) and (3). Here, the correction value is corr, and a 1 and a 2 is a constant. corr is calculated, for example, by equations (4) and (5) and depends on the panel inclination angle (θ) and snow accumulation (SC h The threshold value can be determined based on the amount of snow on the PV panel 70, such as the amount of snow on the PV panel 70 (hereinafter referred to as "snow cover"). In formulas (4) and (5), the larger the corr value, the smaller the threshold value, since the greater the influence of temperature and solar radiation. The threshold values ​​for other weather forecast data can also be corrected using the same calculation method as formulas (2) to (5). T'=T-corr T ×T (2) R' = R-corr R ×R (3) corr T =a 1 θ+a 2 SC h (4) corr R =a 1 θ+a 2 SC h (5)

[0109] The threshold determined as described above makes it possible to reflect the snow accumulation tendency of each PV panel 70 when predicting snow accumulation. In this embodiment, the method in which the threshold determination unit 25 calculates the threshold used in predicting snow accumulation has been described, but the threshold for predicting snow accumulation can also be determined by a similar method.

[0110] FIG. 10 is a flow chart showing a process in which the photovoltaic power generation management device 1 according to the third embodiment predicts snow accretion by threshold value judgment.

[0111] As described above, the threshold determination unit 25 determines a threshold using the snow accumulation record data, weather record data, snow accumulation record data, and equipment information data at the time points corresponding to these record data. The snow accumulation prediction unit 23 compares the determined threshold with the snow accumulation prediction result at the prediction target time point, the weather prediction data at the prediction target time point, and the equipment information data, and outputs the snow accumulation prediction result at the prediction target time point by threshold determination.

[0112] According to this embodiment, the photovoltaic power generation management device 1 uses actual snow accumulation values ​​and the like to determine a threshold value for determining snow accretion on the PV panel 70. This allows the photovoltaic power generation management device 1 to perform highly accurate snow accretion prediction even when the amount of learning data for the snow accretion learning model is small.

[0113] Furthermore, according to this embodiment, when determining the threshold value, the photovoltaic power generation management device 1 corrects the threshold value of the weather forecast data by using the inclination angle information of the PV panel 70, which is the equipment information data. This allows the photovoltaic power generation management device 1 to introduce a threshold value that reflects the snow accretion tendency of each PV panel 70 into the snow accretion prediction.

[0114] (Fourth embodiment) FIG. 11 is a hardware configuration diagram of a photovoltaic power generation management device 1 in the fourth embodiment.

[0115] 11 includes a processor 52 such as a CPU, a main memory device 53 such as a RAM, an auxiliary memory device 54 such as a HDD, a network interface 55 such as a LAN (Local Area Network) board, a device interface 56 such as a memory slot or a memory port, and a bus 57 that connects these devices to each other. The photovoltaic power generation management device 1 is, for example, a computer such as a PC, and includes an input unit 4 such as a keyboard and a mouse, and an output unit 5 such as an LCD.

[0116] In this embodiment, a program for causing a computer to execute information processing of the photovoltaic power generation management device 1 is installed in the auxiliary storage device 54. The photovoltaic power generation management device 1 deploys this program in the main storage device 53 and executes it by the processor 52. This realizes the functions of each block shown in FIG. 2 in the photovoltaic power generation management device 1, and makes it possible to predict the amount of power generation described in the first embodiment. Note that data generated by this information processing is temporarily held in the main storage device 53, or stored and saved in the auxiliary storage device 54.

[0117] This program can be installed, for example, by attaching an external device 58 on which this program is recorded to the device interface 56 and storing this program from the external device 58 in the auxiliary storage device 54. Examples of the external device 58 are a computer-readable recording medium and a recording device incorporating such a recording medium. Examples of the recording medium are a CD-ROM (Compact Disk Read Only Memory), a CD-R (Compact Disk Recordable), a flexible disk, a DVD-ROM (Digital Versatile Disk Read Only Memory), and a DVD-R (Digital Versatile Disk Recordable), and an example of the recording device is a HDD. Also, this program can be installed, for example, by downloading this program via the network interface 55.

[0118] According to this embodiment, it is possible to realize the functions of the photovoltaic power generation management device 1 in the first embodiment by software. Similarly, the functions of the photovoltaic power generation management device 1 in the second and third embodiments can also be realized by software.

[0119] Although several embodiments have been described above, these embodiments are presented only as examples and are not intended to limit the scope of the invention. The novel photovoltaic power generation management device 1 and the like described in this specification can be embodied in various other forms. In addition, various omissions, substitutions, and modifications can be made to the forms of the photovoltaic power generation management device 1 and the like described in this specification without departing from the gist of the invention. The appended claims and their equivalents are intended to include such forms and modifications that fall within the scope and gist of the invention. [Explanation of symbols]

[0120] 1: photovoltaic power generation management device, 2: processing unit, 3: storage unit, 4: input unit 5: output unit, 6: communication unit, 10: photovoltaic power generation management system, 21: Acquisition unit, 22: Snowfall prediction unit, 23: Snow accumulation prediction unit, 23': snow accumulation estimation unit, 24: power generation prediction unit, 25: threshold determination unit, 31: Facility information database, 32: Meteorological satellite database, 33: Weather forecast database, 34: Past performance database, 35: PV power generation forecast database, 40: Satellite image server, 41: weather server, 50: network device, 51: network, 52: processor, 53: main memory device, 54: auxiliary memory device, 55: network interface, 56: device interface, 57: bus, 58: external device, 51: network, 60: weather satellite, 61: receiving antenna, 62: satellite network, 70: PV panel, 80: String configuration, 90: Array configuration, 100: Management office, 200: Solar power plant, 231: Snow accumulation classification section

Claims

1. an acquisition unit that acquires photographed image data of a predetermined area including an installation location of the PV panel at a reference time point and weather forecast data for a prediction target time point after the reference time point; a snowfall prediction unit that predicts the snowfall condition of the specified area at the prediction target time based on the photographed image data and the weather forecast data acquired by the acquisition unit; a snow accumulation prediction unit that predicts a snow accumulation state on the PV panel at the prediction target time based on the snow accumulation state prediction result, the acquiring unit confirms whether there is a missing part in the photographed image data at the reference time point or the weather forecast data at the prediction target time point, and if there is a missing part, acquires the photographed image data at a timing earlier than the reference time point or the weather forecast data at a timing earlier than the prediction target time point. Photovoltaic power generation management device.

2. The snowfall prediction unit includes: Further predicting the snowfall condition in a predetermined area before a prediction target time between the reference time point and the prediction target time point; predicting the snow accumulation condition of the specified region at the prediction target time point using a prediction result of the snow accumulation condition of the specified region before the prediction target time point; The photovoltaic power generation management device according to claim 1 .

3. The solar power generation management device of claim 1, wherein the snowfall prediction unit includes a snowfall prediction model that accepts as input an estimated result of the snowfall conditions at the reference time and the weather forecast data, and outputs information indicating the snowfall depth or the presence or absence of snowfall as a predicted result of the snowfall conditions.

4. The solar power generation management device of claim 1, wherein the snow accumulation prediction unit outputs information indicating the depth of snow or the presence or absence of snow as a predicted result of the snow accumulation condition based on a result of comparing the estimated result of the snow accumulation condition at the reference time point and the weather forecast data with a threshold value.

5. The solar power generation management device of claim 1, wherein the snowfall prediction unit includes a snowfall learning model that accepts as input the prediction result of the snow accumulation condition and at least one of the weather forecast data, equipment information data, and time characteristic information relating to the prediction target time point, and outputs information indicating the presence or absence of snowfall or information indicating the probability of snowfall as the prediction result of the snowfall condition.

6. The solar power generation management device of claim 1, wherein the snow accumulation prediction unit outputs information indicating the presence or absence of snow accumulation or information indicating the probability of snow accumulation as a prediction result of the snow accumulation condition based on a result of comparing the prediction result of the snow accumulation condition and at least one of the weather forecast data, equipment information data, and time characteristic information relating to the prediction target time point with a threshold value.

7. The solar power generation management device of claim 5, wherein the snowfall prediction unit determines whether or not there is snowfall on the PV panel at the prediction target time based on the rate of decrease in the actual power generation amount at the prediction target time relative to the predicted power generation amount value when snowfall on the PV panel at the prediction target time is not taken into account, and generates correct answer data in a snowfall learning model based on the result of the determination.

8. A threshold determination unit that determines the threshold value, The solar power generation management device of claim 6, wherein the threshold determination unit includes a threshold determination model that accepts at least one of snow accumulation history data, weather history data, or snow accumulation history data as input and outputs threshold data for predicting the snow accumulation condition.

9. The photovoltaic power generation management device according to claim 8 , wherein the threshold value determination unit corrects the threshold value data of the weather forecast data by using tilt angle information of the PV panel.

10. an acquisition unit acquires photographed image data of a predetermined area including an installation location of the PV panel at a reference time point and weather forecast data for a prediction target time point after the reference time point; a snowfall prediction unit predicts a snowfall condition in the specified area at the prediction target time based on the photographed image data and the weather forecast data acquired by the acquisition unit; a snow accumulation prediction unit predicts a snow accumulation condition on the PV panel at the prediction target time based on the prediction result of the snow accumulation condition; the acquiring unit confirms whether there is a missing part in the photographed image data at the reference time point or the weather forecast data at the prediction target time point, and if there is a missing part, acquires the photographed image data at a timing earlier than the reference time point or the weather forecast data at a timing earlier than the prediction target time point. A solar power generation management method comprising:

11. an acquisition unit acquires photographed image data of a predetermined area including an installation location of the PV panel at a reference time point and weather forecast data for a prediction target time point after the reference time point; a snowfall prediction unit predicts a snowfall condition in the specified area at the prediction target time based on the photographed image data and the weather forecast data acquired by the acquisition unit; a snow accumulation prediction unit predicts a snow accumulation condition on the PV panel at the prediction target time based on the prediction result of the snow accumulation condition; the acquiring unit confirms whether there is a missing part in the photographed image data at the reference time point or the weather forecast data at the prediction target time point, and if there is a missing part, acquires the photographed image data at a timing earlier than the reference time point or the weather forecast data at a timing earlier than the prediction target time point. A photovoltaic power generation management program that causes a computer to execute a photovoltaic power generation management method including the steps of:

Images