JP2020108188A - Pv output prediction support device, pv output prediction device, pv output prediction support method and pv output prediction support program - Google Patents

Abstract

To provide a PV output prediction support device, a PV output prediction device, a PV output prediction support method and a PV output prediction support program, which can improve prediction precision of PV output.SOLUTION: A storage section 12 stores information of a solar radiation distribution on a past day in a target area, a measurement result of photovoltaic power generation output on the past day and a prediction result on the past day for each of a plurality of prediction models used for predicting the photovoltaic power generation output. An extraction section 15 acquires information of the solar radiation distribution on a prediction target day and extracts a similar day having the solar radiation distribution similar to the prediction target day from selection target days included in a prescribed period in the past days. A prediction error acquisition section 16 acquires a prediction error for each prediction model in the similar day extracted by the extraction section 15 on the basis of the measurement result and the prediction result at every prediction model. A notification section notifies the respective prediction errors which the prediction error acquisition section 16 acquires as the prediction errors of the respective prediction models of prediction of the photovoltaic power generation output on the prediction target day.SELECTED DRAWING: Figure 2

Description

The present invention relates to a PV output prediction support device, a PV output prediction device, a PV output prediction support method, and a PV output prediction support program.

In order to realize a low-carbon society and efficiently use energy, it is expected that the number of consumers who use energy devices such as photovoltaic power generation (PV) and storage batteries will increase in the future. In particular, since the implementation of FIT (Feed In Tariff), the introduction of photovoltaic power generation systems (Photovoltaics) into the distribution system has been rapidly progressing.

If a large amount of solar power generation is introduced into the electric power system of an electric power company, it is feared that fluctuations in the output will have a large impact on the supply and demand operations of the electric power company. In such a situation, in order to carry out stable supply and demand operation, the output of solar power generation in areas of various scales, from wide area level of the entire power system to district level of distribution substation units, can be predicted immediately and accurately. Required to do so. Below, the output of photovoltaic power generation is called "PV output."

The amount of solar radiation used to predict the PV output changes due to daily weather changes. Therefore, the amount of solar radiation can be a large error factor in the prediction of PV output. For example, a temporal and spatial prediction error with respect to the front may cause a large deviation in the amount of solar radiation and the timing of its change, which may have a large impact on supply and demand operation.

In order to reduce such a prediction error, there has been proposed a technique that attempts to improve the prediction accuracy by combining a plurality of PV output prediction models. The PV output can be obtained if the actual amount of solar radiation is known. Therefore, the PV output prediction model can be considered as a model of the actual amount of solar radiation in other words. With this technique, it is possible to make a prediction in consideration of the output fluctuations of the PV output upwards and downwards. In addition, the reliability of the prediction result can be evaluated from the variation of each prediction model and the prediction result from the past to the present.

Here, as a weather information prediction technique, satellite images are grouped based on associated weather information and feature information, a similar group is identified from the feature information of the prediction target satellite images, and the weather information associated with the group is specified. There is a conventional technique for predicting the future weather by the method (Patent Document 1). Further, as a PV output prediction technique, there is a conventional technique for estimating the PV output at the present time by using a predetermined conversion formula with respect to data of a similar day on a prediction target day (Patent Document 2). In addition, the past meteorological results and meteorological forecasts for each time zone are acquired and classified into similar time zones. There is a conventional technique for obtaining a prediction formula for calculating a predicted power generation amount from the amount of solar radiation outside the atmosphere in a time zone (Patent Document 3).

JP, 2016-205930, A JP, 2014-200167, A International Publication No. 2017/026010

However, in the conventional technique that handles a plurality of prediction models, it is important to select which model for supply and demand operation when the prediction results vary greatly among the PV output prediction models. Selections are made based on the experience. In the selection based on the experience of the operator, the accuracy of the PV output prediction result may be deteriorated due to a selection error or the like.

In addition, in the above-mentioned conventional techniques, handling of a plurality of PV output prediction models is not considered, and it is difficult to improve the accuracy of PV output prediction by appropriate model selection.

The disclosed technology is made in view of the above, and provides a PV output prediction support apparatus, a PV output prediction apparatus, a PV output prediction support method, and a PV output prediction support program that improve the accuracy of PV output prediction. The purpose is to

In one aspect of the PV output prediction support device, the PV output prediction support device, the PV output prediction support method, and the PV output prediction support program disclosed in the present application, the storage unit stores information on the solar radiation distribution of the past day in the target area, the past. The measurement result of the solar power generation output of each day and each prediction result in the past day when a plurality of prediction models used for prediction of the solar power generation output are used are stored. The extraction unit acquires information on the solar radiation distribution of the prediction target date and extracts similar days having the solar radiation distribution similar to the prediction target date from the selection target days included in a predetermined period of the past days. To do. The prediction error acquisition unit acquires a prediction error for each prediction model on the similar date extracted by the extraction unit, based on the measurement result and each prediction result. The notification unit notifies each of the prediction errors acquired by the prediction error acquisition unit as a prediction error of each of the prediction models for prediction of the photovoltaic power generation output on the prediction target day.

In one aspect, the present invention can improve the accuracy of PV output prediction.

FIG. 1 is an overall configuration diagram of a PV output prediction support system. FIG. 2 is a block diagram of the PV output prediction apparatus according to the first embodiment. FIG. 3 is a diagram showing a classification result based on clustering. FIG. 4 is an example of a diagram showing an error evaluation result of each prediction model. FIG. 5 is a diagram summarizing the evaluation results of each group. FIG. 6 is a diagram for explaining the selection target date. FIG. 7 is a diagram for explaining the group determination on the selection target day. FIG. 8: is a figure for demonstrating calculation of RMSE between the solar radiation amount distribution of a prediction target day, and the solar radiation amount distribution of each group. FIG. 9: is a figure for demonstrating calculation of RMSE between the solar radiation distribution of a prediction target day, and the solar radiation distribution of a comparison target day. FIG. 10: is a figure for demonstrating calculation of RMSE between the solar radiation amount difference distribution of a prediction target day, and the solar radiation amount difference distribution of a comparison target day. FIG. 11: is a figure for demonstrating calculation of the standard deviation of the solar radiation amount of the prediction target day using the solar radiation amount distribution of the prediction start time of the prediction target date, and the average value of the solar radiation amount of the prediction target day. FIG. 12 is a diagram for explaining a cumulus determination method at the time of analysis based on standard deviation. FIG. 13 is a diagram showing the observation points of the meteorological office in the target area. FIG. 14A is a diagram showing a first page of a Web page that provides information on the result of prediction of the amount of solar radiation. FIG. 14B is a diagram showing a result of solar radiation amount prediction on the first page of the Web page. FIG. 15 is a diagram showing the second page of the Web page for providing information on the PV output prediction result. FIG. 16 is a flowchart of the prediction error notification process of each prediction model performed by the PV output prediction device according to the first embodiment. FIG. 17 is a flowchart of the extraction process of similar days. FIG. 18 is a block diagram of the PV output prediction device according to the second embodiment. FIG. 19 is a diagram illustrating an example of a Web page provided by the PV output prediction device according to the third embodiment. FIG. 20 is a diagram illustrating an example of another web page provided by the PV output prediction apparatus according to the third embodiment. FIG. 21 is a hardware configuration diagram of the PV output prediction device.

Hereinafter, embodiments of the PV output prediction support device, the PV output prediction support device, the PV output prediction support method, and the PV output prediction support program disclosed in the present application will be described in detail with reference to the drawings. The PV output prediction support apparatus, the PV output prediction apparatus, the PV output prediction support method, and the PV output prediction support program disclosed in the present application are not limited to the embodiments described below.

FIG. 1 is an overall configuration diagram of a PV output prediction support system. As shown in FIG. 1, the PV output prediction support system according to the present embodiment includes a PV output prediction device 1, a weather data distribution system 2, and a terminal device 3.

The weather data distribution system 2 acquires, for example, a weather satellite image captured by Himawari-8. Then, the weather data distribution system 2 stores the acquired weather satellite image.

The PV output prediction device 1 acquires a weather satellite image from the weather data distribution system 2. Then, the PV output prediction device 1 calculates the solar radiation amount distribution from the acquired meteorological satellite image, and uses the calculated solar radiation amount distribution to generate information that supports the selection of the prediction model used to predict the PV output. After that, the PV output prediction device 1 provides the terminal device 3 with information for supporting the selection of the generated prediction model. The PV output prediction device 1 provides information that supports selection of a prediction model at intervals of 30 minutes, for example.

Here, the PV output prediction device 1 does not have to directly send the information that supports the selection of the prediction model to the terminal device 3. As a method of providing information that supports selection of a prediction model, for example, the PV output prediction device 1 browses a page on which prediction support information that supports selection of a prediction model is placed on a Web server accessible from the terminal device 3. You may make it possible. In this case, the PV output prediction device 1 updates the information on the Web page at intervals of 30 minutes, for example.

The terminal device 3 is arranged in a power company or the like. The terminal device 3 acquires the prediction support information from the PV output prediction device 1. Then, the terminal device 3 provides the user by outputting the prediction support information to an output device such as a monitor. The user confirms the prediction support information on a monitor or the like and determines which of the plural prediction models is appropriate at the time of prediction among the existing prediction models. Then, the user selects the optimum prediction model and predicts the PV output at the present time.

The creation of the selection support information by the PV output prediction device 1 will be described in detail below. In the present embodiment, the PV output prediction device 1 will be described as a case where PV output is predicted at 30-minute intervals up to 3 hours after the prediction start time, that is, PV output in six time zones. Below, the day when the PV output prediction apparatus 1 makes a prediction is referred to as a “prediction target day”. That is, the prediction target date refers to the day on which the PV output prediction device 1 makes a prediction. In addition, the time at which the prediction is started is referred to as a “prediction start time”, and the time zone targeted for the prediction is referred to as a “prediction target time zone”. In the present embodiment, the prediction target time zone will be described as including a time zone of every 30 minutes with a prediction start time of 9:30 and a prediction end time of 12:30. Further, in the following description, a case where three different prediction models are used as the prediction model will be described. And let each prediction model be a 1st-3rd prediction model. Furthermore, the area where the PV output is predicted is referred to as a “target area”. In this embodiment, the target area is an area including the mainland Kyushu as an example.

FIG. 2 is a block diagram of the PV output prediction apparatus according to the first embodiment. As shown in FIG. 2, the PV output prediction apparatus 1 includes a solar radiation amount distribution estimation unit 11, a storage unit 12, a classification unit 13, a prediction error calculation unit 14, an extraction unit 15, a prediction error acquisition unit 16, a notification unit 17, and a PV. The output prediction unit 18 is included.

The solar radiation distribution estimation unit 11 acquires daily weather satellite images from the weather data distribution system 2. In the present embodiment, the weather satellite image is acquired for each of the prediction target date, the current year until the prediction target date, and all the days of the previous year and the year before last. Below, the day of the year up to the prediction target day, and all the days of the previous year and the year before last are referred to as “past days”.

Next, the solar radiation amount distribution estimation unit 11 obtains the cloud albedo in each predetermined unit area of the target area from the meteorological satellite image in each time zone of each day. Here, the predetermined unit area is, for example, 1 km ² . Further, the solar radiation amount distribution estimation unit 11 calculates the ground surface albedo using the obtained cloud albedo. After that, the solar radiation distribution estimation unit 11 corrects the calculated cloud albedo and ground surface albedo for the correction coefficient of the atmosphere during fine weather, the correction coefficient for considering the cloud type, and the effect of air mass. The total solar radiation is calculated using the coefficient and the horizontal solar radiation incident on the horizontal unit area at the upper end of the atmosphere. Then, the solar radiation amount distribution estimation unit 11 sets the total solar radiation amount in the target area as a solar radiation amount distribution used for PV output prediction. Then, the solar radiation amount distribution estimation unit 11 stores the solar radiation amount distribution in each time zone of each past day in the solar radiation amount distribution database 122.

In addition, the prediction is continuously performed at 10-minute intervals, and the solar radiation distribution estimation unit 11 also acquires the weather satellite image at the prediction start time. Therefore, the solar radiation amount distribution estimation unit 11 also stores the solar radiation amount distribution at the prediction start time of the prediction target day in the solar radiation amount distribution database 122.

Here, the solar radiation amount distribution estimation unit 11 periodically estimates the solar radiation amount distribution in advance in advance, and newly estimates the solar radiation amount distribution with respect to the date and time when the solar radiation amount distribution is not estimated at the timing of performing the PV output prediction process. The distribution may be estimated and added to the solar radiation distribution database 122.

The storage unit 12 has relative humidity data 121, a solar radiation distribution database 122, PV output performance data 123, and PV output prediction result data 124. In the solar radiation amount distribution database 122, the solar radiation amount distribution estimation unit 11 registers the solar radiation amount distribution.

In the relative humidity data 121, the relative humidity for each time zone of the prediction target day and the past day at a predetermined point in the target area acquired from the Meteorological Office of the Meteorological Agency is registered.

As the PV output actual data 123, information on PV output measurement values in the target area for each time zone of the past day is registered. The PV output prediction result data 124 stores the prediction result of the PV output prediction performed by the PV output prediction unit 18 described later. That is, as the PV output prediction result data 124, information on the predicted value of each PV output prediction model in the target area for each time zone of the past day is registered. Then, in the PV output prediction result data 124, three prediction results are registered for each time zone of each past day. The PV output prediction result data 124 also stores the prediction result of the prediction date. The PV output prediction result data 124 also registers the prediction result of the solar radiation amount distribution.

The classification unit 13 acquires the solar radiation amount distribution for each prediction target time zone of the past day from the solar radiation amount distribution database 122. Next, the classification unit 13 classifies the solar radiation distribution for each time zone of the prediction target day of the past day into 16 groups by using the hierarchical clustering without learning. A specific clustering method will be described below.

The classification unit 13 normalizes each amount of solar radiation by the amount of solar radiation outside the atmosphere. Next, the classification unit 13 performs smoothing on the data in each time zone. Next, the classification unit 13 calculates the correlation coefficient σ of the spatial distribution of the amount of solar radiation in the same time zone of each past day by brute force. Next, the classification unit 13 subtracts a value obtained by squaring the correlation coefficient σ from 1, that is, a value obtained by expressing σ of “1-σ ² ”in 1.0 to 0.001 steps as a threshold value. The index of closeness is D. Next, the classification unit 13 combines pairs having a high correlation coefficient. Next, the classification unit 13 calculates the average insolation amount for the combined pair. The classification unit 13 repeats the creation of pairs until the correlation coefficient does not exceed the threshold. Here, the threshold value is, for example, σ=0.3 (D=0.91).

By this clustering, the classification unit 13 can classify the past days into 16 groups as shown in the classification 131 of FIG. FIG. 3 is a diagram showing a classification result based on clustering. The 16 group classifications are standardized solar radiation distributions. Here, in FIG. 3, a diagram created by multiplying the standardized amount of solar radiation for each unit area of each group by the amount of solar radiation outside the atmosphere to obtain the amount of solar radiation and color-coding it according to the size of the amount of solar radiation Represents each of the groups. In the figure, a dark color part is a region with many clouds, and a light color part is a region with few clouds. Next, as shown in the classification 132, the classification unit 13 rearranges the order of the created groups so that the number of clouds increases. As a result, the classification unit 13 determines the groups #1 to #16 shown in the classification 132.

That is, a distribution obtained by multiplying the standardized solar radiation amount distribution of each of the groups #1 to #16 generated by the classification unit 13 by the extra-atmospheric solar radiation amount at a specific date and time is shown by each image of the classification 132 in FIG. It corresponds to the distribution of solar radiation. In the following, the distribution of insolation obtained by multiplying the standardized insolation distribution of groups #1 to #16 by the extra-atmospheric insolation at a specific date and time is simply referred to as “insolation distribution of groups #1 to #16”.

After that, the classification unit 13 outputs the classification result to the prediction error calculation unit 14 and the selection target date narrowing unit 151 of the extraction unit 15. The classification result includes information on past days belonging to each of the groups #1 to #16. Further, the classification unit 13 outputs the representative solar radiation amount distribution of each of the groups #1 to #16 to the selection target date narrowing unit 151.

The prediction error calculation unit 14 receives an input of the classification result from the classification unit 13. Next, the prediction error calculation unit 14 acquires the PV output measurement result for each prediction target time zone of each past day from the PV output actual data 123. In addition, the prediction error calculation unit 14 acquires the PV output prediction result for each prediction target time period of each past day from the PV output prediction result data 124.

Next, the prediction error calculation unit 14 performs error evaluation of each prediction model in each group using RMSE (Root Mean Squared Error) which is the square root of the mean squared error. That is, the prediction error calculation unit 14 performs the error evaluation using the following mathematical expression (1). Here, the target period used for the error evaluation is 3 months including 1 month before and after the target month including the target date for prediction. Here, the reason for narrowing the target period to 3 months, which is 1 month before and after the target month, is that if it is not close to the target date for prediction, a large difference may occur in the prediction results due to differences in seasons and the like. This is because. The RMSE is an example of the “error evaluation function”.

Here, PV _model is a prediction result of PV output. Further, PV _obs is a measurement result of PV output. Here, the square of the difference between the PV _model and PV _obs of the number of days in the target period belonging to the group in which the error evaluation is performed is calculated, and the sum of the calculation results is obtained. N is the number of days within the target period belonging to the group in which the error evaluation is performed.

That is, the prediction error calculation unit 14 takes the square root of the value obtained by dividing the sum of the squared values of the difference between the PV output prediction result and the PV output measurement result on each day of the target period by the number of days of the target period. , RMSE between the predicted PV output value and the actual PV output value is obtained. Then, the prediction error calculation unit 14 creates a prediction error graph representing the error evaluation result of each prediction model using the calculated RMSE. However, in the present embodiment, when the number of samples for which the error evaluation is performed is three or less, the prediction error calculation unit 14 does not perform the error evaluation because it is difficult to perform the appropriate error evaluation.

The prediction error calculation unit 14 performs the error evaluation of each prediction model described above for each prediction target time zone. That is, in this embodiment, the error evaluation of each prediction model in the six time zones included from 9:30, which is the prediction start time, to 12:30, which is 180 minutes ahead, is performed. The prediction error calculation unit 14 creates the prediction error graph illustrated in FIG.

FIG. 4 is an example of a diagram showing an error evaluation result of each prediction model. The prediction error graphs 41 and 42 represent the error evaluation results in any two of the groups #1 to #16. In the prediction error graphs 41 and 42, the vertical axis represents the value of REMS and the horizontal axis represents the time zone. The line segment 21 in the prediction error graphs 41 and 42 represents the error in each time zone of the first prediction model. The line segment 22 represents the error in each time zone of the second prediction model. The line segment 23 represents the error in each time zone of the third prediction model. The smaller the value of RMSE, the smaller the error. Therefore, in the prediction error graph 41, the second prediction model represented by the line segment 22 has the smallest error. In the prediction error graph 42, the third model represented by the line segment 23 has the smallest error.

Further, the prediction error calculation unit 14 writes the Cluster Map indicating the prediction start time and the classification number in the prediction error graphs 41 and 42. In addition, the prediction error calculation unit 14 describes the number of days in the target period belonging to the group represented by the prediction error graph 41, that is, the number of samples used for error evaluation. Here, in the present embodiment, the prediction error calculation unit 14 determines that the number of samples is less than 10 as the number of samples is small and the accuracy of the evaluation result of the error evaluation may be low. The highlighting 24 for notifying is performed.

FIG. 5 is a diagram summarizing the evaluation results of each group. The prediction error calculation unit 14 performs the error evaluation for each prediction model described above for each of the groups #1 to #16, and generates prediction error graphs 201 to 216. Here, the prediction error graphs 201 to 216 are evaluation results for the groups #1 to #16, respectively. As shown in the prediction error graphs 203, 207, 210, and 212 to 216, since the number of samples is small for the groups #3, #7, #10, and #12 to #16, the error evaluation is not performed. In this way, there are groups that do not occur or rarely occur depending on the season of the forecast target day.

The extraction unit 15 performs extraction from similar past days having similar weather conditions such as the solar radiation distribution, the cloud distribution, and the relative humidity on the prediction target day. The extraction unit 15 includes a selection target date narrowing unit 151, a first difference calculation unit 152, a second difference calculation unit 153, a relative humidity difference calculation unit 154, a cloud distribution calculation unit 155, and a similar date extraction unit 156.

As shown in FIG. 6, the selection target date narrowing down unit 151 specifies 45 days before and after the prediction target date and 45 days before the prediction target date from the past days as the selection target days. FIG. 6 is a diagram for explaining the selection target date. Day 31 is the prediction target day. 45 days before and after the same day as the prediction target day and 45 days before the prediction target day are examples of the “predetermined period”. Then, the day 32 is the same day as the prediction target day one year ago. Further, the day 33 is the same day as the prediction target day two years ago. The selection target date narrowing unit 151 identifies the days included in the periods D1, D2, and D3 as the selection target dates.

Next, the selection target date narrowing down unit 151 acquires the solar radiation amount distribution at the predicted start time of the selection target date from the solar radiation amount distribution database 122. The selection target date narrowing unit 151 also receives an input of the classification result from the classification unit 13. Then, the selection target date narrowing down unit 151 determines to which of the groups #1 to #16 the solar radiation distribution of the prediction start time of each selection target date belongs. Here, the group to which the solar radiation distribution of the predicted start time on a specific day belongs is referred to as a “group of specific days”.

For example, the group determination by the selection target date narrowing unit 151 will be described with reference to FIG. 7 as an example. FIG. 7 is a diagram for explaining the group determination on the selection target day. A case will be described in which the eight solar radiation distributions 51 to 58 shown in FIG. 7 are included in the solar radiation distribution at the predicted start time of the selection target day. The selection target date narrowing unit 151 determines from the classification result acquired from the classification unit 13 that the solar radiation distribution 51 belongs to the group #4. Further, the selection target date narrowing unit 151 determines that the solar radiation amount distribution 52 belongs to the group #7. Further, the selection target date narrowing unit 151 determines that the solar radiation amount distribution 53 belongs to the group #16. Further, the selection target date narrowing unit 151 determines that the solar radiation amount distribution 54 belongs to the group #6. In addition, the selection target date narrowing unit 151 determines that the solar radiation amount distribution 55 belongs to the group #8. Further, the selection target date narrowing unit 151 determines that the solar radiation amount distribution 56 belongs to the group #16. In addition, the selection target date narrowing unit 151 determines that the solar radiation amount distribution 57 belongs to the group #15. In addition, the selection target date narrowing-down unit 151 determines that the solar radiation amount distribution 58 belongs to the group #15. In this way, the selection target date narrowing down unit 151 repeats the determination with respect to the distribution of the solar radiation amounts at the prediction start times of all the selection target days.

Next, the selection target date narrowing down unit 151 acquires the solar radiation amount distribution at the prediction start time of the prediction target date from the solar radiation amount distribution database 122. In addition, the selection target date narrowing unit 151 acquires the solar radiation amount distributions of the groups #1 to #16 from the solar radiation amount distribution database 122. Then, the selection target day narrowing unit 151 uses the pair of mathematical expressions (2) to calculate the RMSE between the solar radiation distribution at the prediction start time of the prediction target day and the solar radiation distribution of each of the groups #1 to #16. Ask.

FIG. 8: is a figure for demonstrating calculation of RMSE between the solar radiation amount distribution of a prediction target day, and the solar radiation amount distribution of each group. Here, RMSE in FIG. 8 is obtained by the above-described mathematical expression 2. The distribution of the amount of solar radiation at the prediction start time of the prediction target day has data of the amount of solar radiation for each predetermined unit area. The solar radiation amount distribution map 61 of FIG. 8 is a diagram showing the solar radiation amount distribution in the target region at the prediction start time of the prediction target day. The points 62 arranged in a grid in the solar radiation amount distribution map 61 represent the solar radiation amount for each predetermined unit area. 321 predetermined unit areas are arranged in the other small area in the column direction and 553 are arranged in the row direction, that is, the predetermined unit area after 321×535 is included in the target area, but in FIG. Nine predetermined unit areas are arranged in the column direction and nine are arranged in the row direction. Here, the case where 9×9 predetermined unit areas are included will be described as an example.

Numbers are assigned to each line so that each time the line goes up, the number is incremented by 1 with the point at the upper left lower end as the origin toward the diagram of the solar radiation distribution map 61. Also, the columns are numbered so that each time the column moves to the right, the number increases by one. In the following, the number of each column is represented by i (i≧1). Further, the number of each row is represented by j (j≧1). That is, each point 62 of the solar radiation amount distribution map 61 is expressed as (i, j) with the point at the lower left end facing the paper surface as (1, 1). In FIG. 8, nine predetermined unit areas are arranged in the column direction in the target area and nine predetermined unit areas are arranged in the row direction. That is, the target area includes 9×9 predetermined unit areas.

Further, the solar radiation amount distribution of the group #k (1≦k≦16) also has the solar radiation amount data for each predetermined unit area. The solar radiation amount distribution diagram 63 of FIG. 8 is a diagram showing the solar radiation amount distribution of the group #k (1≦k≦16). The points 64 arranged in a grid in the solar radiation amount distribution map 63 represent the solar radiation amount for each predetermined unit area. Also in this case, each point 64 of the solar radiation amount distribution map 63 is expressed as (i, j) with the point at the lower left end facing the paper surface as (1, 1).

Then, the selection target date narrowing down unit 151 obtains the RMSE between the solar radiation amount distribution at the prediction start time of the prediction target date and the solar radiation amount distribution of each group #1 to #16 by using the above-described mathematical expression (2). ..

Here, the prediction target day (i, j) is the amount of solar radiation at the point 62 represented by (i, j) in the solar radiation amount distribution map 61. The group #k(i,j) is the amount of solar radiation at the point 64 represented by (i,j) in the solar radiation amount distribution map 63. Then, nx*ny is the number of points 62 included in the solar radiation amount distribution map 61. That is, nx is the maximum value of j and ny is the maximum value of i. Here, nx*ny is 9×9.

That is, the selection target day narrowing unit 151 obtains a value obtained by squaring the difference between each point 62 in the solar radiation amount distribution map 61 and each point 64 in the solar radiation amount distribution map 63 for all combinations of (i, j). .. Then, the selection target date narrowing down unit 151 obtains a square root of a value obtained by dividing the total sum of the obtained results by the number of predetermined unit areas included in the target area, and calculates the solar radiation distribution of the prediction start time of each prediction target day and each RMSE, which is the mean square error between the representative solar radiation amounts of groups #1 to #16, is calculated.

Next, the selection target date narrowing down unit 151 determines a group having three RMSEs in ascending order of the RMSE calculated using the mathematical expression (2) to have a solar radiation distribution similar to the prediction start time of the prediction target day. A group having is selected as a similar group. For example, the selection target date narrowing unit 151 selects groups #6, 7, and 4 as similar groups. Here, although three groups are selected in this embodiment, the number of groups is not limited to three. The selection target date narrowing unit 151 may select a smaller number of groups in order to reduce the subsequent similar date selection processing, and to improve the accuracy of the similar day selection processing, a larger number of groups may be selected. It is preferable to select a group of

Next, the selection target date narrowing down unit 151 identifies the selection target dates included in the selected similar group by using the information of the group to which each selection target date determined in advance belongs. Hereinafter, the selection target date specified by the selection target date narrowing unit 151 is referred to as a “comparison target date”.

Then, the selection target day narrowing unit 151 outputs the solar radiation amount distribution of the prediction start time of the prediction target date and the solar radiation amount distribution of the prediction start time of the comparison target day to the first difference calculation unit 152. In addition, the selection target date narrowing unit 151 outputs the information of the comparison target date and the prediction target date to the second difference calculation unit 153. In addition, the selection target date narrowing unit 151 outputs the information of the comparison target date and the prediction target date to the relative humidity difference calculation unit 154. Further, the selection target date narrowing down unit 151 outputs the solar radiation amount distribution at the prediction start time of the prediction target day to the cloud distribution calculating unit 155.

The first difference calculation unit 152 receives input of the solar radiation amount distribution of the prediction start time of the prediction target date and the solar radiation amount distribution of the prediction start time of the comparison target day from the selection target date narrowing unit 151. Then, the first difference calculation unit 152 compares the solar radiation amount distribution at the prediction start time and the solar radiation amount distribution at the prediction start time of the comparison target day with the solar radiation distribution at the prediction start time of the prediction target date. The RMSE between the predicted start time of the target date and the distribution of the amount of solar radiation is calculated. Specifically, the first difference calculating unit 152 uses the following mathematical expression (3) to calculate the RMSE between the solar radiation amount distribution of the prediction start time of the prediction target day and the solar radiation amount distribution of the prediction start time of the comparison target day. To calculate.

The calculation of the RMSE between the solar radiation distribution at the prediction start time of the prediction target day and the solar radiation distribution at the prediction start time of the comparison target day will be described in detail below with reference to FIG. 9. FIG. 9: is a figure for demonstrating calculation of RMSE between the solar radiation distribution of a prediction target day, and the solar radiation distribution of a comparison target day.

The solar radiation amount distribution map 71 of FIG. 9 represents the solar radiation amount distribution at the prediction start time of the prediction target day in the target region. The points 72 arranged in a grid in the solar radiation amount distribution map 71 represent the solar radiation amount for each predetermined unit area. Here, each point 72 of the solar radiation amount distribution map 71 is expressed as (i, j) with the point at the lower left corner toward the paper surface being (1, 1).

Further, the solar radiation amount distribution map 73 in FIG. 9 represents the solar radiation amount distribution at the prediction start time of the comparison target day in the target region. The points 74 arranged in a grid in the solar radiation amount distribution map 73 represent the solar radiation amount for each predetermined unit area. Also in this case, each point 74 of the solar radiation amount distribution map 73 is expressed as (i, j) with the point at the lower left end facing the paper surface as (1, 1).

Then, the first difference calculation unit 152 obtains the RMSE between the solar radiation amount distribution of the prediction start time of the prediction target day and the solar radiation amount distribution of the prediction start time of the comparison target day using the above-described mathematical expression (3).

Here, the prediction target day (i, j) is the amount of solar radiation at the point 72 represented by (i, j) in the solar radiation amount distribution map 71. Further, the comparison target date (i, j) is the amount of solar radiation at the point 74 represented by (i, j) in the solar radiation amount distribution map 73. Then, nx*ny is the number of points 72 included in the solar radiation amount distribution map 71. Here, nx*ny is 9×9.

That is, the first difference calculation unit 152 obtains a value obtained by squaring the difference between each point 72 in the solar radiation amount distribution map 71 and each point 74 in the solar radiation amount distribution map 73 for all combinations of (i, j). Then, the first difference calculation unit 152 obtains a square root of a value obtained by dividing the sum of the obtained results by the number of predetermined unit areas included in the target area, and compares the solar radiation distribution at the prediction start time of the prediction target day with the comparison target. RMSE between the solar radiation distribution at the predicted start time of day is calculated. The first difference calculation unit 152 calculates the RMSE of the solar radiation amount distribution for all the comparison target days.

After that, the first difference calculation unit 152 outputs the RMSE of the solar radiation distribution for each comparison target day calculated using the mathematical expression (3) to the similar date extraction unit 156. Here, the RMSE between the solar radiation amount distribution at the prediction start time of the prediction target day and the solar radiation amount at the prediction start time of the comparison target day is calculated as follows: It shows how different it is from the distribution of insolation at time. That is, it can be said that the smaller the RMSE, the more similar the solar radiation distribution at the predicted start time of the comparison target day and the solar radiation distribution at the predicted start time of the prediction target day.

The second difference calculation unit 153 receives input of information on the prediction target date and the comparison target date from the selection target date narrowing unit 151. Then, the second difference calculation unit 153 calculates the solar radiation distribution in the time zone 1 hour before the prediction start time of the prediction target day and the solar radiation quantity 1 hour before and 1 hour after the prediction start time of the comparison target day. The distribution is acquired from the solar radiation distribution database 122. In addition, the second difference calculation unit 153 acquires, from the PV output prediction result data 124, the solar radiation amount distribution in the time zone one hour after the prediction start time of the prediction target day. Here, the solar radiation amount distribution in the time zone one hour after the prediction start time of the prediction target day is a predicted value by the PV output prediction unit 18. Here, since the second difference calculation unit 153 has high prediction accuracy in the past, the second difference calculation unit 153 obtains the difference distribution of the solar radiation amount using the solar radiation amount distributions one hour before and one hour after, so that the calculation including the past is also performed. Another time zone may be used as long as it is a time zone in which a change in the solar radiation amount near the time can be obtained. For example, when the accuracy of the prediction calculation is high, the second difference calculation unit 153 may use the time zones 0 hours before and 2 hours after the prediction start time for calculating the solar radiation distribution.

Next, the second difference calculation unit 153 subtracts the solar radiation distribution 1 hour before the prediction start time of the prediction target day from the solar radiation distribution 1 hour after the prediction start time of the prediction target day to obtain the prediction target day. Generate the difference distribution of solar radiation. In addition, the second difference calculation unit 153 subtracts the solar radiation distribution 1 hour before the prediction start time of the comparison target day from the solar radiation distribution 1 hour after the prediction start time of the comparison target day to obtain the comparison target day. Generate a solar radiation difference distribution.

The second difference calculation unit 153 uses the solar radiation amount difference distribution of the prediction target date and the solar radiation amount difference distribution of each comparison target date to calculate the solar radiation amount difference distribution of the prediction target date and the solar radiation amount difference distribution of each comparison target day. Calculate the RMSE between. Specifically, the second difference calculating unit 153 calculates the RMSE between the solar radiation amount difference distribution on the prediction target date and the solar radiation amount difference distribution on each comparison target date using the following mathematical expression (4).

Hereinafter, the calculation of the RMSE between the solar radiation amount difference distribution on the prediction target date and the solar radiation amount difference distribution on the comparison target date will be described in detail with reference to FIG. 10. FIG. 10: is a figure for demonstrating calculation of RMSE between the solar radiation amount difference distribution of a prediction target day, and the solar radiation amount difference distribution of a comparison target day.

The solar radiation amount difference distribution map 81 of FIG. 10 represents the solar radiation amount difference distribution of the prediction target day in the target region. The points 82 arranged in a grid pattern in the solar radiation amount difference distribution diagram 81 represent the difference between the solar radiation amounts one hour before and one hour before the predicted start time for each predetermined unit area. Here, each point 82 of the solar radiation amount difference distribution map 81 is expressed as (i, j) with the point at the lower left end facing the paper surface as (1, 1).

Further, the solar radiation amount difference distribution map 83 of FIG. 10 represents the solar radiation amount difference distribution of the comparison target day in the target region. The points 84 arranged in a grid in the solar radiation amount difference distribution map 83 represent the difference between the solar radiation amount in the time zone one hour before the predicted start time for each predetermined unit area. Also in this case, each point 84 of the solar radiation amount difference distribution map 83 is represented as (i, j) with the point at the lower left end facing the paper surface as (1, 1).

Then, the second difference calculation unit 153 obtains the RMSE between the solar radiation amount difference distribution on the prediction target date and the solar radiation amount difference distribution on each comparison target date using the above-described mathematical expression (4).

Here, the prediction target daily difference (i, j) is the solar radiation amount difference at the point 82 represented by (i, j) in the solar radiation amount difference distribution map 81. The comparison target difference (i, j) is the amount of solar radiation at the point 84 represented by (i, j) in the solar radiation amount difference distribution map 83. Then, nx*ny is the number of points 82 included in the solar radiation amount difference distribution map 81. Here, nx*ny is 9×9.

In other words, the second difference calculation unit 153 squares the difference between each point 82 in the solar radiation amount difference distribution map 81 and each point 84 in the solar radiation amount difference distribution map 83 for all combinations of (i, j). Ask. Then, the second difference calculation unit 153 obtains a square root of a value obtained by dividing the total sum of the obtained results by the number of predetermined unit regions included in the target region, and calculates the difference in solar radiation difference distribution on the prediction target date and the solar radiation on the comparison target date. RMSE, which is the mean squared error with the quantity difference distribution, is calculated. The second difference calculation unit 153 calculates the RMSE of the solar radiation amount difference distribution for all the comparison target days.

Then, the second difference calculation unit 153 outputs the calculated RMSE of the solar radiation amount difference distribution for each comparison target day to the similar date extraction unit 156. Here, obtaining the difference in the amount of solar radiation corresponds to obtaining the movement of the cloud. Therefore, it is possible to specify the comparison target day in which the movement of the clouds is similar to the prediction target day using the solar radiation amount difference distribution. That is, it can be said that the smaller the RMSE, the more similar the movement of the cloud on the comparison target day to the movement of the cloud on the prediction target day.

The cloud distribution calculation unit 155 receives the input of the solar radiation amount distribution of the prediction start time of the prediction target date and the solar radiation amount distribution of the prediction start time of the comparison target day from the selection target date narrowing unit 151. Next, the cloud distribution calculation unit 155 calculates the average value of the solar radiation amount at the prediction start time of the prediction target day using the solar radiation distribution at the prediction start time of the prediction target day. Further, the cloud distribution calculation unit 155 calculates the average value of the amount of solar radiation at the prediction start time of the comparison target day using the distribution of the amount of solar radiation at the prediction start time of the comparison target day.

Next, the cloud distribution calculation unit 155 obtains the standard deviation of the solar radiation amount on the prediction target day by using the solar radiation amount distribution at the prediction start time of the prediction target day and the average value of the solar radiation amount on the prediction target day. In addition, the cloud distribution calculation unit 155 obtains the standard deviation of the solar radiation amount of the prediction target day using the solar radiation amount distribution of the prediction start time of each comparison target day and the average value of the solar radiation amount of each comparison target day.

The method for obtaining the standard deviation of the amount of solar radiation is the same on both the prediction target day and the comparison target day. Therefore, the calculation of the standard deviation of the solar radiation amount of the prediction target day using the solar radiation amount distribution of the prediction start time of the prediction target date and the average value of the solar radiation amount of the prediction target day will be described in detail below with reference to FIG. 11. explain. FIG. 11: is a figure for demonstrating calculation of the standard deviation of the solar radiation amount of the prediction target day using the solar radiation amount distribution of the prediction start time of the prediction target date, and the average value of the solar radiation amount of the prediction target day.

The solar radiation amount distribution map 91 of FIG. 11 represents the solar radiation amount distribution of the prediction start time of the prediction target day in the target region. The points 92 arranged in a grid in the solar radiation amount distribution map 91 represent the solar radiation amount for each predetermined unit area. Here, each point 92 of the solar radiation amount distribution map 91 is expressed as (i, j) with the point at the lower left corner toward the paper surface being (1, 1).

The cloud distribution calculation unit 155 sums up the values of the points 92 in the solar radiation amount distribution map 91 and divides by the number of points 92 included in the solar radiation amount distribution map 91 to obtain the solar radiation amount at the prediction start time of the prediction target day. Calculate the average value of.

Then, the cloud distribution calculation unit 155 obtains the standard deviation of the solar radiation amount on the prediction target day using the following mathematical expression (5).

Here, the prediction target day (i, j) is the amount of solar radiation at the point 92 represented by (i, j) in the solar radiation amount distribution map 91. Moreover, the prediction target day average is an average value of the amount of solar radiation on the prediction target day. Then, nx*ny is the number of points 92 included in the solar radiation amount distribution map 91. Here, nx*ny is 9×9.

That is, the cloud distribution calculation unit 155 obtains a value obtained by squaring the difference between each point 92 in the solar radiation amount distribution map 91 and the average value of the solar radiation amount for all combinations of (i, j). Then, the cloud distribution calculation unit 155 obtains the square root of the value obtained by dividing the sum of the obtained results by the number of predetermined unit areas included in the target area, and calculates the standard deviation of the solar radiation amount on the prediction target day. The second difference calculation unit 153 similarly calculates the standard deviation of the amount of solar radiation for each comparison target day.

After that, the cloud distribution calculation unit 155 outputs the calculated standard deviation of the amount of solar radiation for the prediction target day and each comparison target day to the similar day extraction unit 156. Here, the standard deviation of the amount of solar radiation can be said to be a value that summarizes the deviation from the average of the amount of solar radiation at each point in the target area. That is, it can be said that the larger the standard deviation is, the larger the deviation of the amount of solar radiation from the average is at many points, and it can be said that a lot of fine clouds are generated. On the other hand, as the standard deviation is smaller, the amount of solar radiation is the same as the average at most points, and it can be said that there is no cloud. That is, it is possible to determine the presence or absence of a fine cumulus cloud at the time of fine weather by using the standard deviation of the amount of solar radiation.

FIG. 12 is a diagram for explaining a cumulus determination method at the time of analysis based on standard deviation. For example, the image 101 is an image when the standard deviation is small. That is, the image 101 represents the case where there is almost no cloud. On the other hand, the image 102 is an image when the standard deviation is large. That is, the image 102 represents a state in which many fine cumulus clouds are generated. It is assumed that there is a state represented by images 103 to 105 as a state to be compared with this. In this case, the comparison state represented by the images 103 to 105 has a small standard deviation. Then, as can be seen from the images 103 to 105, there is almost no cloud. On the other hand, the states represented by the images 106 to 108 have a large standard deviation. Then, as can be seen from the images 106 to 108, many clouds are generated. That is, from the standard deviation, it can be determined that the states of the images 106 to 108 are such that the cumulus state is not similar to the image 101. Further, from the standard deviation, the states of the images 103 to 105 can be determined to be the states in which the cumulus state is not similar to the image 102.

The relative humidity difference calculation unit 154 receives input of information on the comparison target date and the prediction target date from the selection target date narrowing unit 151. Then, the relative humidity difference calculation unit 154 acquires the relative humidity RH at the prediction start time at the predetermined observation point on the comparison target day and the prediction target day from the relative humidity data 121.

Here, the observation points of the meteorological office that observe the relative humidity of the target area used in this example are eight points 111 to 118 shown in FIG. FIG. 13 is a diagram showing the observation points of the meteorological office in the target area. In the present embodiment, the relative humidity difference calculation unit 154 acquires the relative humidity at the prediction start time of the comparison target day and the prediction target day at the points 111 to 118.

Next, the relative humidity difference calculation unit 154 calculates RMSE, which is the root mean square error of the relative humidity RH (Relative Humidity) of the comparison target day and the prediction target day, using the following mathematical expression (6).

Here, the RH target day time (r) is the relative humidity RH of each of the eight points r (1≦r≦8) on the comparison target day. The RH prediction target day (r) is the relative humidity RH of each of the eight points r (1≦r≦8) on the prediction target day. Then, the square of the difference between the RH comparison target date (r) and the RH prediction target date (r) is calculated for each of the eight points and summed. M is the number of observation points. In this embodiment, M=8.

That is, the relative humidity difference calculation unit 154 obtains a value obtained by squaring the difference between the relative humidity RH of the prediction target day and the relative humidity RH of the comparison target day at each of the points 111 to 118 for all the observation points. Then, the relative humidity difference calculation unit 154 calculates the square root of the value obtained by dividing the total sum of the calculated results by the number of observation points, and calculates the RMSE of the relative humidity RH between the comparison target date and the prediction target date. The relative humidity difference calculation unit 154 calculates the RMSE of the relative humidity RH for all the comparison target days.

After that, the relative humidity difference calculation unit 154 outputs the calculated relative humidity RMSE between the comparison target date and the predicted target date to the similar date extraction unit 156. Here, the larger the value of the RMSE of the relative humidity, the more different the humidity condition between the comparison target day and the prediction target day. That is, it can be seen that the comparison target day having a large RMSE of the relative humidity RH has a significantly different humidity state from the prediction target day.

The similar day extraction unit 156 receives, from the first difference calculation unit 152, the input of the RMSE of the solar radiation distribution for each comparison target day calculated using the mathematical expression (3). The similar date extraction unit 156 also receives, from the second difference calculation unit 153, the input of the RMSE of the solar radiation amount difference distribution for each comparison target day calculated using the mathematical expression (4). Further, the similar day extraction unit 156 receives from the cloud distribution calculation unit 155 the input of the standard deviation of the amount of solar radiation for the prediction target date and each comparison target date calculated using the mathematical expression (5). Further, the similar date extraction unit 156 receives, from the relative humidity difference calculation unit 154, an input of the RMSE of the relative humidity between the comparison target date and the prediction target date calculated using the mathematical expression (6).

Next, the similar day extraction unit 156 obtains the product of the RMSE of the solar radiation amount distribution and the RMSE of the solar radiation amount difference distribution for each comparison target day. Then, the similar date extraction unit 156 determines the rank in the ascending order of the value of the calculation result. That is, the similar day extraction unit 156 ranks the comparison target days in which the distribution of insolation and the moving state of clouds are more similar to each other.

Next, the similar day extraction unit 156 obtains the difference in the standard deviation of the amount of solar radiation from the prediction target day for each comparison target day. Then, the similar date extraction unit 156 excludes, from the ranked comparison target dates, the comparison target dates whose standard deviation of the comparison target date is outside the predetermined range with respect to the standard deviation of the prediction target date. That is, the similar date extracting unit 156 excludes days from which the cumulus state is significantly different from the prediction target date from the ranked comparison target days.

For example, under sunny conditions, the similar day extraction unit 156 performs the following prediction target day exclusion processing. The similar day extraction unit 156 excludes comparison target days outside the range of ±0.015 if the standard deviation of the standardized solar radiation amount on the prediction target date is less than 0.05. Further, the similar day extraction unit 156 excludes the comparison target days outside the range of ±0.030 if the standard deviation of the standardized solar radiation amount on the prediction target date is 0.05 or more and less than 0.01. Further, the similar date extraction unit 156 excludes comparison target days outside the range of ±0.050 if the standard deviation of the standardized solar radiation amount on the prediction target date is 0.0.10 or more.

Next, the similar date extraction unit 156 removes, from the ranked comparison target dates, the comparison target dates that are larger than the designated value of the RMES of the relative humidity RH. That is, the similar date extraction unit 156 excludes days from which the humidity state is significantly different from the prediction target date, from the ranked comparison target days.

For example, in the present embodiment, the similar date extraction unit 156 excludes the comparison target dates having a relative humidity RMSE of greater than 10% from the ranked comparison target dates. However, it is preferable to change this condition depending on the number of comparison days to be obtained. For example, if the condition is that the number of samples to be used for comparison is 20 or more, if the conditions are not satisfied when the comparison date for which the relative humidity RMSE is greater than 10% is set as the exclusion date, the RMSE It is preferable that the threshold is set to 20% or 30% and the conditions to be excluded from the comparison target day are relaxed. However, when changing the threshold value, it is preferable to determine the maximum value of the threshold value, for example, it is preferable to determine the maximum value of the threshold value to be 30%.

Next, the similar date extraction unit 156 raises the ranks of the remaining comparison days to be removed and determines the ranks. After that, the similar date extraction unit 156 extracts the top three comparison target dates as similar days. Then, the similar date extraction unit 156 outputs the extracted information of the similar date to the prediction error acquisition unit 16.

The prediction error acquisition unit 16 receives an input of information on a similar date for the prediction target date from the similar date extraction unit 156 of the extraction unit 15. Then, the prediction error acquisition unit 16 acquires the error evaluation result of each prediction model of the group to which the designated similar day belongs from the prediction error calculation unit 14. After that, the prediction error acquisition unit 16 outputs the acquired error evaluation result of each prediction model on each similar day to the notification unit 17.

The PV output prediction unit 18 has first to third prediction models used for PV output prediction in advance. Then, the PV output prediction unit 18 acquires the information necessary for the prediction from the weather data distribution system 2. Then, the PV output prediction unit 18 uses each prediction model to predict the PV output in each prediction target time zone of the prediction target day in each case. Then, the PV output prediction unit 18 outputs the prediction result of the PV output when each prediction model is used to the notification unit 17. The PV output prediction unit 18 stores the solar radiation amount forecast distribution calculated in the PV output forecast calculation in the solar radiation amount distribution database 122. The predicted solar radiation distribution is overwritten with the solar radiation distribution of the solar radiation distribution estimation unit 11 at the same time after the prediction.

The notification unit 17 receives the input of the error evaluation result of each prediction model on each similar day from the prediction error acquisition unit 16. In addition, the notification unit 17 acquires the solar radiation amount distribution at the prediction target start time of the extracted similar days from the solar radiation amount distribution database 122. Further, the notification unit 17 acquires the PV output measurement result on the similar day from the PV output performance data 123. In addition, the notification unit 17 acquires the PV output prediction result when each prediction model on the similar day is used from the PV output prediction result data 124. Furthermore, the notification unit 17 receives, from the PV output prediction unit 18, the input of the prediction result of the PV output of the prediction target day when each prediction model is used. Then, the notification unit 17 calculates the solar radiation distribution of the prediction target start time of the similar day, the PV output measurement result on the similar day, and the prediction models of the PV output, the error evaluation result of each prediction model on the similar day, and The prediction result of the PV output of the prediction target day when each prediction model is used is transmitted to the terminal device 3 to notify the user.

An example of information provision by the notification unit 17 when a web page is used will be described below with reference to FIGS. 14A, 14B, and 15. FIG. 14A is a diagram showing a first page of a Web page that provides information on the result of prediction of the amount of solar radiation. FIG. 14B is a diagram showing the result of solar radiation amount prediction on the first page of the Web page. Further, FIG. 15 is a diagram showing a second page of the Web page for providing information on the PV output prediction result.

As illustrated in FIG. 14A, the web page 220 displays an image 221 that represents the distribution of solar radiation at the prediction start time of the prediction target day. Further, on the web page 220, images 222 to 224 showing the distribution of the amount of solar radiation at the predicted start times of the three similar days are displayed. Further, on the web page 220, an image 225 showing the average solar radiation amount on the prediction target day and the prediction result of the solar radiation amount for each prediction model is displayed.

In the image 225, other than the above, for example, as shown in FIG. 14B, the observation result, a continuous model based on the observation value of the average of 10 minutes before the prediction start time, a sunny model assuming a sunny day, and a 10-minute interval are performed. It is also possible to display the predicted value to be executed, the past actual value of the predicted value to be executed at intervals of 10 minutes, the predicted value based on the short-term precipitation forecast, and the like. FIG. 14B shows an image 225 showing an example of performing prediction calculation up to 6 hours ahead at 10 minute intervals. Graph 2251 represents the observation result. Also, the solid line portion of graph 2252 represents the estimated value, and the dashed line portion represents the persistence model based on the estimated value of the previous 10-minute average. A graph 2253 represents a clear weather model assuming a clear time. Also, the graph 2254 represents the predicted value at 30-minute intervals. Further, the graph 2255 is a predicted value based on the short-term forecast for precipitation by the Japan Meteorological Agency.

For the similar days 222 to 223, images 226 to 228 representing the prediction results of the solar radiation amount on the similar days and the solar radiation amount for each prediction model are displayed. In addition, the web page 220 may display information on groups #1 to #16 that are similar to the predicted start time zone of the predicted start date.

The user can refer to the web page 220 displayed on the terminal device 3 to determine the degree of similarity of the solar radiation distribution between the prediction target date and each similar date from the images 221 to 224. Further, the user refers to the Web page 220 displayed on the terminal device 3 and displays an image 225 of how the average of the solar radiation amounts in the eight meteorological offices in the target area on the target day for prediction is predicted. It can be understood from ~228. Furthermore, the user can grasp from the images 225 to 228 how much the predicted result of the solar radiation amount of each prediction model is actually different from the measured result in the past prediction on each similar day.

Further, when the user selects the display of the second page, the web page 230 shown in FIG. 15 is displayed on the terminal device 3. On the web page 230, images 231 to 233 showing the evaluation results of each similar day are displayed. Further, on the Web page 230, images 234 to 236 showing the PV output measurement result and the prediction result on each similar day are displayed.

By referring to the images 231-233, the user can determine which prediction model is most likely to perform the most appropriate prediction at a date and time similar to the prediction time zone of the prediction target day. In addition, the user refers to the images 234 to 236 to confirm which prediction model actually predicted the PV output at a date and time similar to the prediction time zone of the prediction target day. You can That is, the user can select the prediction model that is considered to be the most accurate in the prediction time zone of the prediction target day by referring to the Web pages 220 and 230.

Next, with reference to FIG. 16, a flow of notification processing of a prediction error of each prediction model by the PV output prediction apparatus 1 according to the present embodiment will be described. FIG. 16 is a flowchart of the prediction error notification process of each prediction model performed by the PV output prediction device according to the first embodiment.

The classification unit 13 acquires the solar radiation amount distribution in the predicted time zone of the past day from the solar radiation amount distribution database 122. Then, the classification unit 13 performs hierarchical clustering to classify the past days into 16 groups 1 to #16 (step S1).

The prediction error calculation unit 14 obtains the PV output measurement result of the prediction target time zone of the past day from the PV output actual data 123, and the PV output prediction result of each prediction model of the past day prediction target time zone. From the PV output prediction result data 124. Next, the prediction error calculation unit 14 evaluates the error of each prediction model for each of the groups #1 to #16 generated by the classification unit 13, and calculates the prediction error for each prediction model in the prediction target time zone of the past day. (Step S2).

The extraction unit 15 extracts the comparison target date from the selection target dates included in the past days. Then, the extraction unit 15 executes a similar day extraction process by using the prediction start time and the prediction start time of the comparison target day and the solar radiation distribution around 1 hour, and the relative humidity of the prediction target day and the comparison target day. (Step S3).

The prediction error acquisition unit 16 acquires the prediction error of the similar date extracted by the extraction unit 15 from the prediction error calculation unit 14 (step S4).

The notification unit 17 outputs the prediction error of the similar days acquired by the prediction error acquisition unit 16 to the terminal device 3 to notify the user (step S5). At this time, the notification unit 17 may output a prediction result or the like when each PV output prediction model is used.

Next, with reference to FIG. 17, a flow of extraction processing of the similar date by the extraction unit 15 will be described. FIG. 17 is a flowchart of the extraction process of similar days.

The selection target date narrowing unit 151 acquires the classification result of the solar radiation amount distribution and the solar radiation amount distribution of each group from the classification unit 13. In addition, the selection target date narrowing down unit 151 acquires the solar radiation amount distribution at the predicted start time of the selection target date from the solar radiation amount distribution database 122. Further, the selection target date narrowing unit 151 selects 45 days before and after the prediction target date from the past days as the selection target date. Next, the selection target date narrowing down unit 151 identifies the group to which the distribution of the solar radiation amount of the predicted start time of each selection target day belongs, from the classification result acquired from the classification unit 13 (step S101).

Next, the selection target date narrowing down unit 151 acquires the solar radiation amount distribution at the prediction start time of the prediction target date from the solar radiation amount distribution database 122. Then, the selection target date narrowing down unit 151 calculates the RMSE between the solar radiation distribution at the prediction start time of the prediction target date and the solar radiation distribution of each group (step S102).

Next, the selection target date narrowing down unit 151 sets the top three groups with the smallest RMSE of the solar radiation amount distribution as similar groups. Then, the selection target date narrowing down unit 151 narrows down the comparison target dates by setting the selection target dates belonging to the similar group as the comparison target dates (step S103).

The first difference calculation unit 152 acquires the solar radiation amount distribution of the prediction start time of the prediction target date and the comparison target date from the selection target date narrowing unit 151. Then, the first difference calculation unit 152 calculates the RMSE of the solar radiation amount distribution between the prediction target day and the prediction start time of the comparison target day using the mathematical expression (3) (step S104).

The second difference calculation unit 153 acquires information on the prediction target date and the comparison target date from the selection target date narrowing unit 151. Then, the second difference calculation unit 153 acquires, from the insolation distribution database 122, the insolation distribution in the time zone about one hour before and after the prediction start time of the prediction target day and the comparison target day. Here, the solar radiation amount distribution one hour after the prediction start time of the prediction target day is a predicted value. Next, the second difference calculation unit 153 calculates the difference between the solar radiation amount distribution in the time zone 1 hour before the predicted start time and the predicted start time of each of the comparison target dates and the solar radiation amount distribution in the time zone 1 hour after. Then, the difference distribution of insolation is obtained. Next, the second difference calculation unit 153 calculates the RMSE between the solar radiation amount difference distribution of the prediction target date and the solar radiation amount difference distribution of each target date by using the mathematical expression (4), and moves the cloud area. A speed difference is calculated (step S105).

The cloud distribution calculation unit 155 acquires the solar radiation amount distribution of the prediction start time of the prediction target date and the comparison target date from the selection target date narrowing unit 151. Next, the cloud distribution calculation unit 155 calculates an average value of the solar radiation amounts at the prediction start times of the prediction target day and the comparison target day from the solar radiation amount distributions of the prediction start time of the prediction target day and the comparison target day. Next, the cloud distribution calculation unit 155 uses the average value of the solar radiation amount distribution and the solar radiation amount at the prediction start time of the prediction target day and the comparison target day in the formula (5) to calculate each of the prediction target day and the comparison target day. The respective cloud distributions are obtained by obtaining the standard deviation of the amount of solar radiation (step S106).

The relative humidity difference calculation unit 154 acquires information on the prediction target date and the comparison target date from the selection target date narrowing unit 151. Next, the relative humidity difference calculation unit 154 acquires the relative humidity RH at each of the predetermined observation points at the prediction start time of the prediction target day and the comparison target day from the relative humidity data 121. Then, the relative humidity difference calculation unit 154 calculates RMSE between the relative humidity RH at the measurement start time of the prediction target day and the relative humidity RH at the measurement start time of the comparison target day using the mathematical expression (6) (step). S107).

The similar day extraction unit 156 receives the input of the RMSE of the solar radiation amount distribution for each comparison target day from the first difference calculation unit 152. The similar date extraction unit 156 also receives, from the second difference calculation unit 153, the input of the RMSE of the solar radiation amount difference distribution for each comparison target day. The similar day extraction unit 156 also receives, from the cloud distribution calculation unit 155, an input of the standard deviation of the amount of solar radiation for the prediction target date and each comparison target date. Further, the similar date extraction unit 156 receives the input of the relative humidity RMSE between the comparison target date and the prediction target date from the relative humidity difference calculation unit 154. Then, the similar day extraction unit 156 obtains the product of the RMSE of the solar radiation distribution and the RMSE of the solar radiation distribution for each comparison target day. Then, the similar date extraction unit 156 determines the rank in the ascending order of the value of the calculation result. That is, the similar day extraction unit 156 ranks the comparison target days in which the distribution of insolation and the moving state of the cloud are more similar to each other (step S108).

Next, the similar day extraction unit 156 obtains the difference in the standard deviation of the amount of solar radiation from the prediction target day for each comparison target day. Then, the similar date extraction unit 156 determines a comparison target day having a large difference in value with respect to the calculated standard deviation of the prediction target day as an exclusion day. The similar date extraction unit 156 also determines the comparison target date having a large relative humidity RMSE value as the exclusion date. Then, the similar date extraction unit 156 removes the exclusion date from the ranked comparison target dates (step S109). That is, the similar day extraction unit 156 excludes days in which cumulus states are significantly different from the prediction target days and days in which humidity states are significantly different from the prediction target days from the ranked comparison target days.

The similar date extraction unit 156 again descends the ranks of the comparison target days excluding the excluded days from the top to determine the ranks. Then, the similar date extraction unit 156 determines the top three comparison target dates of the confirmed rank as the similar dates of the prediction target date (step S110).

After that, the similar date extraction unit 156 notifies the determined similar date to the prediction error acquisition unit 16 (step S111).

As described above, the PV output prediction apparatus according to the present embodiment classifies past days into groups having common characteristics from the solar radiation distribution and calculates the prediction error of each prediction model for each group. Then, a similar day having a solar radiation environment similar to the prediction target day is specified, and the user is notified of the prediction error of each prediction model of the group to which the similar day belongs. Thereby, the user can determine the prediction model used for the prediction of the PV output from the prediction error obtained by the rational calculation, instead of relying on the experience. That is, the PV output prediction apparatus according to the present embodiment can contribute to improvement in PV output prediction accuracy.

Further, the PV output prediction apparatus according to the present embodiment identifies similar days by using the solar radiation amount distribution, cloud movement state, cloud distribution, and relative humidity RH as the solar radiation environment. In this way, by identifying many similar conditions that affect the PV output and identifying similar days, it is possible to improve the accuracy of identifying similar days when the PV output is similar to the prediction target date. The output prediction accuracy can be improved.

Furthermore, the PV output prediction apparatus according to the present embodiment extracts the selection target date of the same season as the prediction target date from the past days when identifying the similar days, and further, the solar radiation distribution from the selection target days. Narrow down the comparison target dates that belong to the group similar to the predicted date. This makes it possible to reduce the number of comparison targets when identifying similar days, reduce the processing load, and identify similar days at high speed.

Further, the PV output predicting apparatus according to the present embodiment provides the user with a large amount of information such as the distribution of solar radiation amount on the prediction target day and the similar days, the solar radiation amount prediction result, the error evaluation result for each prediction model, and the PV output prediction result. To provide. As a result, the user can evaluate the prediction model from various angles and can select a more appropriate prediction model. That is, the PV output prediction apparatus according to the present embodiment can also contribute to the improvement of the PV output prediction accuracy in terms of the type of information to be provided.

Here, in the present embodiment, the PV output prediction apparatus 1 having the PV output prediction unit 18 has been described, but another device is provided with the function of the PV output prediction unit 18, and similar day extraction or prediction error of similar days is prevented. The PV output prediction support device may have the remaining functions such as acquisition. Even in that case, the PV output prediction support device can contribute to the improvement of the PV output prediction accuracy by providing the user with the prediction error of the similar days. That is, the PV output prediction device 1 corresponds to an example of the “PV prediction support device”.

Further, in the present embodiment, the PV output prediction apparatus 1 has the solar radiation amount distribution estimation unit 11, but other devices are provided with the function of the solar radiation amount distribution estimation unit 11, and the information on the solar radiation amount distribution obtained by the device is provided. May be acquired and used by the PV output prediction device 1.

FIG. 18 is a block diagram of the PV output prediction device according to the second embodiment. The PV output prediction apparatus 1 according to the present embodiment has a usage prediction model determination unit 19 in addition to the units of the first embodiment. The PV output prediction apparatus 1 according to the present embodiment is different from the first embodiment in that a prediction model used for PV output prediction is automatically determined. In the following description, unless otherwise specified, each unit having the same reference numeral as that in FIG. 2 has the same function and will not be described.

The use prediction model determination unit 19 acquires, from the prediction error acquisition unit 16, the similar date and the result of the error evaluation of the prediction model on each similar date. Next, the use prediction model determination unit 19 determines the rank of each prediction model for each prediction time period on each similar day using the result of the error evaluation. That is, the use prediction model determination unit 19 sets the prediction model in descending order of the error evaluation result in a certain time zone.

Next, the usage prediction model determination unit 19 gives a score to the rank of each prediction model of each similar day that is the top three comparison target days. The higher the rank of the error evaluation result, the higher the score. In addition, weighting is performed on the upper comparison target days, and the total value of the scores of the prediction models on the top three comparison target days is calculated. Then, the usage prediction model determination unit 19 determines the prediction model with the highest value of the calculated specific total value as the usage prediction model. After that, the usage prediction model determination unit 19 notifies the notification unit 17 of the determined usage prediction model.

The notification unit 17 receives the notification of the usage prediction model from the usage prediction model determination unit 19. Further, the notification unit 17 acquires, from the PV output prediction result data 124, the PV output prediction result at the prediction start time of the prediction target day, which is performed using the specified usage prediction model. Then, the notification unit 17 transmits the prediction result of the PV output at the prediction start time of the prediction target day, which is performed using the usage prediction model, to the terminal device 3 and notifies the user of the result.

Here, in the present embodiment, the PV output prediction apparatus 1 notifies the user of the prediction result using the determined usage prediction model, but in addition to this, the prediction result using another prediction model is used as a reference. You may notify.

As described above, the PV output prediction apparatus according to the present embodiment determines the prediction model used for prediction from the result of the error evaluation, and provides the user with the PV output prediction result using the determined prediction model. To do. As a result, the PV output prediction apparatus according to the present embodiment can acquire the PV output prediction result using the rationally selected prediction model, and can improve the PV output prediction accuracy. In addition, the user can obtain the PV output prediction result using a reasonably selected prediction model that is considered to be appropriate, and can appropriately control the PV output.

Next, a third embodiment will be described. The PV output prediction apparatus according to the present embodiment differs from the first embodiment in that it further provides information for selecting a prediction model. The PV output prediction apparatus according to this embodiment is also represented by the block diagram of FIG. In the following description, unless otherwise specified, each unit having the same reference numeral as that in FIG. 2 has the same function and will not be described.

The notification unit 17 acquires, from the insolation distribution database 122, the insolation distribution of each time period included within 3 hours before and after the prediction start time of the prediction target day and the similar day. As shown in FIG. 19, the notification unit 17 generates a Web page that displays an image 303 representing the solar radiation amount distribution of the prediction start time and the prediction start time of the similar day. FIG. 19 is a diagram illustrating an example of a Web page provided by the PV output prediction device according to the third embodiment.

In addition, the notification unit 17 causes the generated Web page to display a time selection button 301 for displaying an image of a predetermined time before and after the predicted start time. Further, the notification unit 17 displays a slide bar 302 for continuously changing the time on the Web page. Then, the notification unit 17 displays the generated Web page on the terminal device 3 and provides the information to the user.

Then, when the user selects any one of the time selection buttons 301 using the input device of the terminal device 3, the notification unit 17 displays the solar radiation distribution map of the time zone corresponding to the selected time selection button 301 as an image. It is displayed as 303. When the user uses the input device of the terminal device 3 to select continuous times from a specific time with the slide bar 302, the notification unit 17 continues from the solar radiation distribution in the time zone corresponding to the selected time. The image 303 is switched and displayed on the distribution of the amount of solar radiation in the time zone.

In addition, when the notification unit 17 receives an instruction to provide more detailed information for selecting a prediction model using the input device of the terminal device 3, the notification unit 17 is similar in each time zone of a predetermined period including the prediction time zone. The extraction unit 15 is requested to extract the date. After that, the notification unit 17 acquires from the extraction unit 15 the information of the similar days and the information of the group to which the similar days belong in each time zone of the predetermined period from the extraction unit 15. Then, as shown in FIG. 20, the notification unit 17 is a Web page that displays similar date information 311 indicating information on similar days and group information 312 indicating information on a group to which each similar date belongs in each time period of a predetermined period. Is displayed on the terminal device 3 and provided to the user. FIG. 20 is a diagram illustrating an example of another web page provided by the PV output prediction apparatus according to the third embodiment.

By using the time selection button 301 and the slide bar 302, the user can grasp the change in the distribution of insolation in the vicinity of the predicted time zone, and can further confirm the movement of the cloud from the change in distribution of insolation. Thereby, the user can select a more appropriate prediction model by predicting a change in the weather from the movement of clouds near the prediction time zone and using the change in the prediction model.

In addition, the user obtains information about similar days in each time zone of a predetermined period including the estimated time zone and the group to which those similar days belong, and thereby the appearance rate of similar days and the appearance of groups near the estimated time zone. The rate can be grasped, the similarity rate for each similar day or the prediction target day of the group can be confirmed, and a more appropriate prediction model can be selected.

As described above, the PV output prediction apparatus according to the present embodiment notifies the user of the movement of clouds near the predicted time zone, the similar date, and the appearance status of the group. As a result, the user can select a more appropriate prediction model, and thus the PV output prediction apparatus according to the present embodiment can contribute to the improvement of the PV output prediction accuracy.

(Hardware configuration)
Next, with reference to FIG. 21, a hardware configuration of the PV output prediction apparatus 1 described in each of the above embodiments will be described. FIG. 21 is a hardware configuration diagram of the PV output prediction device.

As shown in FIG. 21, the PV output prediction device 1 includes a CPU (Central Processing Unit) 901, a memory 902, a hard disk 903, and a network interface 904. The CPU 901 is connected to the memory 902, the hard disk 903, and the network interface 904 by a bus.

The hard disk 903 is an auxiliary storage device and realizes the function of the storage unit 12. Further, the hard disk 903 has the solar radiation distribution estimation unit 11, the classification unit 13, the prediction error calculation unit 14, the extraction unit 15, the prediction error acquisition unit 16, the notification unit 17, and the PV output prediction illustrated in FIGS. Various programs including a program that realizes the functions of the unit 18 and the usage prediction model determination unit 19 are stored. The network interface 904 is an interface for communication with the weather data distribution system 2 and the terminal device 3.

The CPU 901 reads out various programs from the hard disk 903, expands them on the memory 902, and executes the programs to execute the insolation distribution estimation unit 11, the classification unit 13, the prediction error calculation unit 14, and the extraction illustrated in FIGS. The functions of the unit 15, the prediction error acquisition unit 16, the notification unit 17, the PV output prediction unit 18, and the usage prediction model determination unit 19 are realized.

The solar radiation distribution estimation unit 11, the classification unit 13, the prediction error calculation unit 14, the extraction unit 15, the prediction error acquisition unit 16, the notification unit 17, the PV output prediction unit 18, and the use prediction model determination unit of each of the above-described embodiments. The programs for realizing the respective functions of 19 are not necessarily stored in the hard disk 903 from the beginning.

For example, a "portable physical medium" such as a flexible disk (FD), a CD-ROM (Compact Disk-Read Only Memory), a DVD, a magneto-optical disk, an IC (Integrated Circuit) card, which is inserted into the PV output prediction device 1. The program is stored in. Then, the PV output prediction device 1 may read out the program from these and execute it.

Furthermore, the program is stored in “another computer (or server)” or the like connected to the PV output prediction apparatus 1 via a public line, the Internet, a LAN (Local Area Network), a WAN (Wide Area Network), or the like. deep. Then, the PV output prediction device 1 may read out the program from these and execute it.

1 PV Output Prediction Device 2 Meteorological Data Distribution System 3 Terminal Device 11 Solar Radiation Distribution Estimate Part 12 Storage Part 13 Classification Part 14 Prediction Error Calculation Part 15 Extraction Part 16 Prediction Error Acquisition Part 17 Notification Part 18 PV Output Prediction Part 19 Usage Prediction Model Determination unit 121 Relative humidity data 122 Solar radiation amount distribution database 123 PV output actual result data 124 PV output prediction result data 151 Selection target date narrowing unit 152 First difference calculation unit 153 Second difference calculation unit 154 Relative humidity difference calculation unit 155 Cloud distribution Calculation unit 156 Similar day extraction unit

Claims (10)

A storage unit that stores information on the solar radiation distribution of the past day in the target area, the measurement result of the solar power generation output of the past day, and the prediction result of the past day for each of a plurality of prediction models used for prediction of the solar power generation output. When,
An extraction unit that obtains information on the solar radiation distribution of the prediction target date, and extracts similar days having the solar radiation distribution similar to the prediction target date from the selection target days included in a predetermined period among the past days. ,
A prediction error acquisition unit that acquires a prediction error for each prediction model on the similar day extracted by the extraction unit, based on the measurement result and the prediction result for each prediction model,
A PV output prediction, comprising: a notification unit that notifies each of the prediction errors acquired by the prediction error acquisition unit as a prediction error of each of the prediction models for prediction of the photovoltaic power generation output on the prediction target day. Support device. A classification unit that classifies the past days into a plurality of groups based on the solar radiation distribution,
A prediction error calculation unit that acquires the prediction result for each prediction model for each group, and calculates the prediction error for each prediction model for each group,
The prediction error acquisition unit acquires the prediction error for each of the prediction models calculated by the prediction error calculation unit for the group to which the similar day extracted by the extraction unit belongs. The PV output prediction support device described in 1. The PV output prediction support apparatus according to claim 2, wherein the prediction error calculation unit calculates the prediction error using an error evaluation function based on the measurement result and the prediction result of the past day. The extraction unit identifies the group to which each of the selection target days belongs based on the solar radiation distribution of the selection target date, and the prediction target date for each of the groups based on the solar radiation distribution of the prediction target date. The PV output prediction support apparatus according to claim 2 or 3, wherein a similar date is obtained from the selection target dates that belong to a predetermined number of groups with a small error. The extraction unit obtains the forecast target day and the predetermined meteorological state of each of the selection target days based on the solar radiation distribution of the selection target date, and the prediction target date and each of the selection target days The PV output prediction support apparatus according to any one of claims 1 to 4, wherein the similar day is extracted based on the distribution of the amount of solar radiation and the state of the predetermined weather. The extraction unit is
A first difference calculation unit that calculates a first difference between the solar radiation distribution of the prediction target date and the solar radiation distribution of each of the selection target days;
The moving speed of the cloud area of the prediction target day is calculated from the solar radiation distribution of the prediction target day, the moving speed of the cloud area of each of the selection target days is calculated from the solar radiation distribution of each of the selection target days, and the prediction target day is calculated. A second difference calculation unit that calculates a respective second difference between the moving speed of the cloud area and the moving speed of the cloud area on each of the selection target days,
The cloud distribution of the prediction target day is calculated based on the solar radiation distribution of the prediction target day, and the cloud distribution of each of the selection target days is calculated based on the solar radiation distribution of each of the selection target days. A cloud distribution difference acquisition unit that calculates a difference in the cloud distribution with each of the selection target dates,
A humidity difference acquisition unit that acquires the relative humidity of each of the selection target dates and the prediction target days, and calculates the difference of the relative humidity between each of the selection target days and the prediction target date,
On the basis of the first difference and the second difference, from among the days excluding the days on which the difference in the cloud distribution is larger than a first predetermined value and the days on which the difference in the relative humidity is larger than a second predetermined value from the selection target day The PV output prediction support apparatus according to any one of claims 1 to 5, further comprising: a similar day extraction unit that extracts the similar days. Based on each of the prediction errors acquired by the prediction error acquisition unit, a usage prediction model determination unit that determines a usage model used for prediction of the photovoltaic power generation output of the prediction target day from the prediction model is further provided. The PV output prediction support device according to any one of claims 1 to 6. A storage unit that stores the information of the solar radiation distribution of the past day in the target area, the measurement result of the solar power output of the past day, and the prediction result of the past day for each of a plurality of prediction models used for prediction of the solar power output. ,
An information extraction unit that obtains information on the solar radiation distribution of the prediction target date, and extracts similar days having a solar radiation distribution similar to the prediction target date from the selection target days included in a predetermined period among the past days,
A prediction error acquisition unit that obtains a prediction error for each prediction model on the similar day extracted by the extraction unit, based on the measurement result and the prediction result for each prediction model,
Based on the prediction error acquired by the prediction error acquisition unit, a determination unit that determines the use model used for prediction of the photovoltaic power generation output of the prediction target day from the prediction model,
And a prediction unit that predicts a photovoltaic power generation output by using the use model determined by the determination unit. Information on the solar radiation distribution of the past day in the target area, obtain the measurement result of the solar power output of the past day and the prediction result of the past day for each of a plurality of prediction models used to predict the solar power output,
Obtaining information on the solar radiation distribution of the prediction target date, extracting a similar day having a solar radiation distribution similar to the prediction target date from the selection target days included in a predetermined period among the past days,
Based on the measurement result on the past day and the prediction result for each prediction model, obtain the prediction error for each prediction model on the extracted similar day,
The PV output prediction support method, wherein each of the obtained prediction errors is notified as a prediction error of each of the prediction models for prediction of the photovoltaic power generation output on the prediction target day. Information on the solar radiation distribution of the past day in the target area, obtain the measurement result of the solar power output of the past day and the prediction result of the past day for each of a plurality of prediction models used to predict the solar power output,
Obtaining information on the solar radiation distribution of the prediction target date, extracting similar days having a similar solar radiation distribution from the past days,
Based on the measurement result on the past day and the prediction result for each prediction model, obtain the prediction error for each prediction model on the selected similar day,
A PV output prediction support program, which notifies each of the obtained prediction errors as a prediction error of each of the prediction models for prediction of the photovoltaic power generation output on the prediction target day.