Solar power generation forecasting system and solar power generation forecasting method

Abstract

To calculate a power generation amount prediction value of a photovoltaic power generation facility in a short period of time ahead from a prediction time point.SOLUTION: A photovoltaic power generation amount prediction system includes: a storage device which stores the past power generation amount actual value of a photovoltaic power generation facility, and the past weather information in association with date and time information; and a prediction device which calculates a power generation amount prediction value of the photovoltaic power generation facility in each frame in a prediction object period starting from the next frame of a first frame being a frame including the current time point. The prediction device calculates the power generation amount prediction value of each frame in the prediction object period by using the power generation amount actual value or power generation amount prediction value in an immediately-previous frame when the first frame is the frame subsequent to the third frame after the prescribed number from the second frame including the time of sunrise, specifies the past date that was the weather equivalent to the weather in the prediction object period as the weather-equivalent day when the first frame is the frame prior to the third frame, and calculates the power generation amount prediction value of each frame in the prediction object period by using the power generation amount actual value of each frame in the same time zone as the prediction object period in the weather-equivalent day.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a solar power generation prediction system and a solar power generation prediction method that calculates a predicted value of the amount of power generated in a short time from the prediction point in time by a solar power generation facility. [Background technology]

[0002] To ensure a stable supply of high-quality electricity, power companies control the output of generators in various locations every day to balance the supply and demand of electricity.

[0003] For example, power companies use the power generation operation plan determined based on power demand forecasts created by the previous day to perform EDC (Economical Load Dispatching Control) and LFC (Load Frequency Control) daily, combining multiple power demand forecasts for several minutes to tens of minutes or several hours ahead, and precisely control the output of their power generators (see, for example, Patent Document 1). [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2013-062953 [Overview of the project] [Problems that the invention aims to solve]

[0005] In recent years, solar power generation facilities, whose output fluctuates significantly depending on the amount of sunlight, have rapidly become widespread, impacting the balance of electricity supply and demand. Therefore, various technologies for predicting the power output of solar power generation facilities are being developed.

[0006] For example, a technology has been developed to predict the amount of electricity generated by a solar power plant by estimating the solar radiation intensity at a given date and time using cloud movement vectors obtained from images of the sky. However, this technology requires the installation of observation equipment for sky photography at each location of the solar power plant, and while it can predict the amount of electricity generated in the very short term, such as a few seconds to a few minutes, it is difficult to predict the amount of electricity generated several tens of minutes in advance.

[0007] Another method uses a sustained model that assumes the current power generation level will remain unchanged for some time, allowing for predictions of power generation levels several tens of minutes in advance. However, the predicted power generation levels obtained using this method are simply delayed versions of past power generation levels, making it impossible to predict power generation levels during the dawn period. Furthermore, it cannot account for patterns of change such as increasing or decreasing trends in power generation, resulting in time delays in the predicted power generation levels and limiting the accuracy of the predictions.

[0008] For these reasons, there is a need for technology that can more accurately predict the amount of electricity generated by solar power generation facilities a predetermined time in advance, such as a few minutes to a few hours ahead, including the time of dawn.

[0009] The present invention has been made in view of the above problems, and one of its objectives is to provide a solar power generation prediction system and a solar power generation prediction method that can more accurately predict the amount of power generated by a solar power generation facility a predetermined time in advance from the prediction point, for example, a short time in advance of a few minutes to a few hours. [Means for solving the problem]

[0010] One of the present inventions for achieving the above objective is a solar power generation forecasting system for calculating predicted power generation values ​​of a solar power generation facility, comprising: a storage device that stores past power generation performance values ​​of the solar power generation facility and past weather information including weather and meteorological data in association with date and time information; and a forecasting device that calculates predicted power generation values ​​of the solar power generation facility in each frame within a forecast target period, which is configured by dividing a day into fixed time intervals, and starts from the frame following the first frame which is the frame including the present time, wherein the forecasting device calculates predicted power generation values ​​of the solar power generation facility in each frame, which is configured by dividing a day into fixed time intervals, and the forecasting device calculates predicted power generation values ​​of the solar power generation facility in each frame, which starts from the frame following the first frame which is the frame including the present time, If the frame is the third frame or later, a predetermined number of frames, starting from the second frame which includes the time of departure, a first process is performed to calculate the predicted power generation value for each frame within the forecast period using the actual power generation value or predicted power generation value for the immediately preceding frame. If the first frame is a frame before the third frame, a past date with weather corresponding to the weather of the forecast period is identified as a weather equivalent day, and a second process is performed to calculate the predicted power generation value for each frame within the forecast period using the actual power generation value for each frame in the same time period as the forecast period on the weather equivalent day.

[0011] According to the solar power generation prediction system of the present invention, it is possible to accurately calculate predicted power generation values ​​for solar power generation facilities a short period of time, such as a few minutes to a few hours, from a prediction point that includes the time of dawn.

[0012] Another aspect of the present invention for achieving the above objective is a solar power generation forecasting system, wherein in the second process, if the number of past dates identified as weather equivalent days is greater than or equal to a predetermined number, the forecasting device calculates a predicted power generation value for each time slot within the forecasting period, after applying a predetermined weight using fuzzy inference to the actual power generation values ​​for each time slot in the same period on each weather equivalent day, according to the degree of similarity between the weather data for the forecasting period and the weather data for each weather equivalent day.

[0013] According to the solar power generation prediction system of the present invention, since a predetermined number or more past dates can be identified as weather-corresponding days, it becomes possible to accurately calculate the predicted power generation value of solar power generation facilities a short time from the prediction point in time using fuzzy inference.

[0014] Another aspect of the present invention for achieving the above objective is a solar power generation forecasting system, wherein in the second process, if the number of past dates identified as weather equivalent days is less than a predetermined number, the forecasting device adds past dates with weather predetermined according to the type of weather during the forecast period as weather equivalent days until the number of weather equivalent days reaches the predetermined number, and calculates the forecast value of power generation for each time slot within the forecast period by multiplying the actual power generation value for each time slot in the same time period on each weather equivalent day by a multiplier determined according to the degree of similarity between the weather during the forecast period and the weather on each weather equivalent day.

[0015] According to the solar power generation prediction system of the present invention, even when it is not possible to identify a predetermined number of past dates as weather-corresponding days, it is possible to accurately calculate the predicted power generation value of a solar power generation facility a short time from the prediction point in time.

[0016] Another aspect of the present invention for achieving the above objective is a solar power generation prediction system, wherein in the second process, if at least the last frame of the prediction period is the third frame or later, the prediction device performs the second process for frames before the third frame within the prediction period, and performs the first process for frames that are the third frame or later within the prediction period.

[0017] According to the solar power generation prediction system of the present invention, it is possible to accurately calculate the predicted power generation value of a solar power generation facility a short time from the prediction point, regardless of whether at least the last frame of the prediction period falls before or after the third frame.

[0018] Another aspect of the present invention for achieving the above objective is a solar power generation forecasting system, which includes a first ratio calculation device that calculates a first ratio, which is the rate of change of the external horizontal solar radiation from the previous frame, for each frame by dividing the value of the external horizontal solar radiation at the installation site of the solar power generation equipment in each frame by the value in the previous frame, wherein the forecasting device calculates the predicted power generation value for each frame within the forecast period by multiplying the actual power generation value or predicted power generation value in the previous frame by the first ratio in the first processing.

[0019] According to the solar power generation prediction system of the present invention, it is possible to accurately calculate the predicted power generation value of a solar power generation facility a short time from the prediction point in time using a first ratio.

[0020] Another aspect of the present invention for achieving the above objective is a solar power generation forecasting system, which includes a second ratio calculation device that calculates a second ratio for each frame, which is the rate of change of the calculated value of the solar power generation equipment from the previous frame, by dividing the calculated value of the solar power generation equipment, which is calculated for each frame using a predetermined calculation formula, by the calculated value in the previous frame, wherein in the first process, the forecasting device calculates the predicted value of the power generation amount for each frame within the forecasting period by multiplying the actual value of the power generation amount or the predicted value of the power generation amount in the previous frame by the second ratio.

[0021] According to the solar power generation prediction system of the present invention, it is possible to accurately calculate the predicted power generation value of a solar power generation facility a short time from the prediction point in time using a second ratio.

[0022] Furthermore, the problems disclosed in this application and their solutions will be made clear from the description in the section on embodiments for carrying out the invention and from the drawings. [Effects of the Invention]

[0023] According to the present invention, it is possible to accurately calculate the predicted power generation amount of a solar power generation facility a short time from the prediction point in time. [Brief explanation of the drawing]

[0024] [Figure 1] This figure shows an example of the configuration of a power system 1, including a solar power generation forecasting system 100. [Figure 2] This figure shows an example of the configuration of the solar power generation forecasting system 100. [Figure 3] This figure shows an example of information stored in DB131 in the past. [Figure 4] This figure shows an example of information about the solar power generation equipment 200 stored in the equipment DB132. [Figure 5] This figure shows an example of the information on the first and second ratios stored in the ratio DB133. [Figure 6] This figure shows an example of the information stored in the search priority DB134. [Figure 7] This figure shows an example of the daily maximum value of extra-atmospheric horizontal solar radiation for the past year, calculated by the ratio calculation unit 121A. [Figure 8] This figure shows an example of the maximum daily power generation amount from the solar power generation equipment 200 over the past year, calculated by the ratio calculation unit 121A'. [Figure 9] This figure shows an example of the daily ratio of horizontal solar radiation outside the atmosphere (solid line) and the daily ratio of power generation (dashed line). [Figure 10] This figure shows an example of the search period that the historical data search unit 121C searches. [Figure 11] This figure shows another example of the search period that the historical data search unit 121C searches. [Figure 12] This diagram shows the algorithm for calculating a power generation performance correction value PVactT, which is a weighted value based on temperature applied to the actual power generation performance value PVact(i). [Figure 13] This diagram shows the algorithm for calculating a power generation performance correction value PVactW, which is a weighted value based on wind speed applied to the actual power generation performance value PVact(i). [Figure 14]This figure shows the algorithm for calculating the power generation performance correction value PVactSR, which is a weighted value based on solar radiation applied to the actual power generation performance value PVact(i). [Figure 15] This diagram shows the algorithm for calculating the predicted power generation value PVactfore from three power generation performance correction values ​​PVactT, PVactW, and PVactSR. [Figure 16] This figure shows other examples of membership functions used when calculating the power generation performance correction values ​​PVactT, PVactW, and PVactSR. [Figure 17] This diagram shows the algorithm for calculating the predicted power generation value PVfore. [Figure 18] This diagram shows the algorithm by which the power generation prediction unit 121D calculates a predicted power generation value based on the actual power generation value generated by the solar power generation equipment 200 in n-1 timeframes. [Figure 19] This flowchart shows an example of how the solar power generation forecasting system 100 operates when calculating a predicted power generation value for a short period of time from the current forecast point in the solar power generation facility 200. [Figure 20] Figure 19 is a flowchart showing a specific example of step S1040. [Figure 21] This flowchart shows another example of how the solar power generation forecasting system 100 operates when calculating a predicted power generation value for a short period of time from the forecast point in time at the solar power generation facility 200. [Modes for carrying out the invention]

[0025] The following matters will become clear from this specification and the accompanying drawings. The present invention will be described below with reference to the accompanying drawings in accordance with one embodiment thereof. In this embodiment, identical or similar components may be denoted by the same reference numeral and their descriptions may be omitted.

[0026] Figure 1 is a diagram showing an example of the configuration of a power system 1 including a solar power generation prediction system 100 according to this embodiment.

[0027] The power system 1 consists of a solar power generation forecasting system 100, solar power generation equipment 200, weather observation meter 300, and Japan Meteorological Agency DB (database) 400. The solar power generation forecasting system 100, solar power generation equipment 200, weather observation meter 300, and Japan Meteorological Agency DB 400 are connected in a state where they can communicate via a communication network 500. The communication network 500 is, for example, a LAN (Local Area Network), WAN (Wide Area Network), dedicated line, power line communication network, various public communication networks, etc.

[0028] The solar power generation equipment 200 includes a solar power generation panel 210, which is constructed using, for example, polycrystalline silicon power generation elements, monocrystalline silicon power generation elements, thin-film power generation elements, etc. The solar power generation equipment 200 also includes a power conditioner (PCS: Power Conditioning Subsystem) 220 that converts the DC generated by the solar power generation panel 210 into AC and supplies it to the power grid 600, but an inverter may be provided instead of the power conditioner 220. The solar power generation equipment 200 transmits a power generation actual value, which indicates the amount of electricity actually generated by the solar power generation equipment 200, to the solar power generation forecasting system 100 via the communication network 500, as one of the physical quantities related to electricity.

[0029] The weather observation meter 300 is a measuring instrument that measures various physical quantities related to the weather in the area where the solar power generation facility 200 is installed. In this embodiment, the weather observation meter 300 includes, for example, a pyranometer for observing the amount of solar radiation in the area where the solar power generation facility 200 is installed, a thermometer for observing the temperature in the same area, and an anemometer for observing the wind speed in the same area. The weather observation meter 300 transmits to the solar power generation forecasting system 100, via the communication network 500, weather data including actual values ​​such as solar radiation information observed by the pyranometer, temperature information observed by the thermometer, and wind speed information observed by the anemometer, as well as weather information including the weather (sunny, rainy, etc.). The weather information transmitted from the weather observation meter 300 to the solar power generation forecasting system 100 may be either a digital signal or an analog signal. If the weather information is an analog signal, the weather information is converted digitally by the solar power generation forecasting system 100 so that the solar power generation forecasting system 100 can perform digital signal processing on it.

[0030] The Japan Meteorological Agency DB400 is a database system that has the function of transmitting various types of weather information (weather, weather data (solar radiation, temperature, wind speed, etc.)). In response to a transmission request from the solar power generation forecasting system 100, the Japan Meteorological Agency DB400 transmits forecast values ​​and historical values ​​of solar radiation, temperature, wind speed, and weather for a specified date and time to the solar power generation forecasting system 100.

[0031] The solar power generation forecasting system 100 is a system that calculates predicted power generation values ​​for the solar power generation equipment 200 at regular intervals (in this embodiment, every 30 minutes as an example) that include a short period of time (for example, several tens of minutes to several hours) from sunrise on the day the forecast is performed to predict the amount of power generated by the solar power generation equipment 200.

[0032] Figure 2 shows an example of the configuration of the solar power generation prediction system 100 according to this embodiment.

[0033] The solar power generation forecasting system 100 is comprised of a computer including an input device 110, a control device 120, a storage device 130, and a communication device 140, as means of realizing the above functions.

[0034] The solar power generation forecasting system 100 acquires various information via the input device 110. Specifically, the solar power generation forecasting system 100 acquires actual meteorological information such as solar radiation, temperature, and wind speed observed by the weather observation meter 300 via the input device 110. The solar power generation forecasting system 100 also acquires actual power generation values, which indicate the amount of power actually generated by the solar power generation equipment 200, via the input device 110. Furthermore, the solar power generation forecasting system 100 acquires meteorological information for the area where the solar power generation equipment 200 is installed, predicted at the time of the forecast execution on the forecast execution date (current time), from the Japan Meteorological Agency DB 400 via the input device 110. Note that the solar power generation forecasting system 100 may be configured to acquire solar radiation, temperature, wind speed, etc., from the Japan Meteorological Agency DB 400 instead of from the weather observation meter 300.

[0035] The control device 120 is composed of an arithmetic processing unit 121 such as a CPU (Central Processing Unit) or MPU (Micro Processing Unit), and a memory 122 such as RAM (Random Access Memory).

[0036] The arithmetic processing unit 121 implements various functions for predicting the amount of power generated by the solar power generation equipment 200 by executing a program stored in the storage device 130. This program may be stored on a recording medium such as a CD-ROM, or it may be distributed via a communication network 500 and installed in the solar power generation prediction system 100.

[0037] Memory 122 temporarily stores various information necessary for calculations when the arithmetic processing unit 121 is executing processing based on the above program.

[0038] The storage device 130 is composed of a storage device such as a hard disk and stores various programs necessary for executing processing in the control device 120, as well as data necessary for executing various programs. In this embodiment, the storage device 130 includes a past DB 131, an equipment DB 132, a ratio DB 133, and a search priority DB 134. In this embodiment, these DBs are represented, for example, as relational databases which are table sets having columns and rows.

[0039] Figure 3 shows an example of information stored in the past DB131. The past DB131 has fields that store information such as the date and time, the installation location of the solar power generation equipment 200 (latitude and longitude), the amount of solar radiation observed by the solar radiation meter of the weather station 300, the temperature observed by the thermometer of the weather station 300, the wind speed observed by the anemometer of the weather station 300, the actual amount of power generated from the solar power generation equipment 200, and weather information, using a time frame that is constructed by dividing the day into fixed time intervals (e.g., 30 minutes). Note that one time frame can be interpreted as a 30-minute time interval.

[0040] Figure 4 shows an example of information about the solar power generation equipment 200 stored in the equipment DB 132. The equipment DB 132 has fields that store information such as the installation location of the solar power generation equipment 200, the installation orientation angle, tilt angle, type (polycrystalline silicon type power generation element, monocrystalline silicon type power generation element, thin film type power generation element, etc.), capacity, and the capacity of the power conditioner 220. If there are multiple solar power generation equipment 200 for which the solar power generation amount prediction system 100 is to calculate the power generation amount prediction value, the above information will be stored in the equipment DB 132 for each of the multiple solar power generation equipment 200 that are to be calculated.

[0041] Figure 5 shows the first ratio R, which will be described later, stored in ratio DB133. sr1 ~R srx , second ratio R pv1 ~R pvxAn example of the information is shown. Specifically, the ratio DB133 is the rate of change of the extraterrestrial horizontal solar irradiance at each frame at the installation location of the solar power generation facility 200, obtained for each frame by dividing the value at each frame by the value at the immediately preceding frame, which is the first ratio R sr1 ~R srx and the rate of change of the calculated value of the power generation amount of the solar power generation facility 200 calculated for each frame using a predetermined calculation formula, obtained for each frame by dividing the calculated value at each frame by the calculated value at the immediately preceding frame, which is the second ratio R pv1 ~R pvx and has a field for storing the information indicating them.

[0042] The calculation formula for the power generation amount of the solar power generation facility 200 is, for example, the following formulas (1) to (4).

[0043] Tpa = Ta+(A / (B×V 0.8 +1)+2)×Ga - 2 ···(1) (Tpa: solar cell panel temperature (°C), Ta: outside air temperature (°C), A: coefficient (for example, rooftop type "50"), B: coefficient (for example, rooftop type "0.38"), V: wind speed (m / s), Ga: inclined plane solar irradiance (kW / m2))

[0044] [[ID=*23]] Kpt = 1+αmax×(Tpa - 25)···(2) (Kpt: temperature correction coefficient, αmax: maximum output temperature coefficient (1 / °C), Tpa: solar cell panel temperature (°C))

[0045] System output coefficient = Kloss × Kpt ···(3)

[0046] (Kloss: change over time (dirt, deterioration), wiring resistance loss, inverter loss, etc., a fixed coefficient generally used according to the equipment type, etc. to consider these, Kpt: temperature correction coefficient)

[0047] Calculated value (Pw) = inclined plane solar irradiance (Ga) × system output coefficient × panel capacity···(4) The solar panel temperature Tpa may be obtained from a sensor installed in the photovoltaic power generation equipment 200, or it may be calculated from the ambient temperature Ta. The inclined surface solar radiation Ga can be calculated from the ambient horizontal solar radiation, taking into account attenuation by the atmosphere.

[0048] The search priority DB134 stores the information shown in Figure 6, which will be described later. Specifically, the search priority DB134 has a field that stores information indicating the weather (e.g., sunny, rainy, cloudy, snowy) of n frames that are predicted to be ahead of the predicted execution date, multiple past weather conditions (e.g., sunny, rainy, cloudy, snowy) that have been given priority according to the weather of n frames that are predicted to be ahead of the predicted execution date, and a magnification Xi associated with these multiple past weather conditions.

[0049] The calculation processing unit 121 includes a ratio calculation unit 121A (ratio calculation unit 121A'), a frame determination unit 121B, a past data search unit 121C, and a power generation prediction unit 121D (prediction device) as functions for the solar power generation prediction system 100 to predict the amount of power generated by the solar power generation equipment 200.

[0050] In this embodiment, the solar power generation prediction system 100 can calculate predicted power generation values ​​for the solar power generation equipment 200 for 48 time slots per day, for example, with each 30-minute interval being considered as one time slot.

[0051] The ratio calculation unit 121A (first ratio calculation device) calculates the amount of horizontal solar radiation outside the atmosphere for one year at the location where the solar power generation equipment 200 is installed. Specifically, the ratio calculation unit 121A calculates the amount of horizontal solar radiation outside the atmosphere for 48 frames per day, going back one year. The past year here may be the year up to one year before the prediction execution date, or it may be the period of a past fiscal year prior to the prediction execution date (April 1st to March 31st of the following year). In this embodiment, the past year is assumed to be the period of a past fiscal year prior to the prediction execution date. Horizontal solar radiation outside the atmosphere is the solar constant (1.367 kW / m²), which is the radiant energy from the sun received outside the Earth's atmosphere. 2This is the horizontal component of ) and is calculated by specifying the latitude and longitude of the installation site of the solar power generation equipment 200 and the date and time. Here, sunlight is scattered and absorbed as it passes through the atmosphere, so the sunlight that reaches the ground surface within the atmosphere is attenuated, and this amount of attenuation changes depending on the atmospheric conditions and the wavelength of sunlight. Therefore, in this embodiment, the horizontal solar radiation outside the atmosphere, which is not affected by the atmospheric conditions and the wavelength of sunlight, is adopted. When the ratio calculation unit 121A has completed the calculation of the horizontal solar radiation outside the atmosphere for the past year, the first ratio R sr1 ~R srx This calculates the horizontal solar radiation outside the atmosphere for frame t at the prediction time, and frame t+1 (frame Δt (30 minutes) ahead of frame t), respectively. sr (t),E sr Let (t+Δt) be the first ratio of the horizontal solar radiation outside the atmosphere, R sr1 ~R srx is, E sr (t+Δt) / E sr (t) is calculated. The first ratio R for this year sr1 ~R srx This is stored in the ratio DB133 within the storage device 130. Also, the first ratio R for one year sr1 ~R srx By calculating this, the first ratio R for any time of year can be determined. sr1 ~R srx It also becomes possible to use [this method].

[0052] The ratio calculation unit 121A' (second ratio calculation device) calculates the amount of horizontal solar radiation outside the atmosphere for one year at the location where the solar power generation equipment 200 is installed. Specifically, the ratio calculation unit 121A' calculates the amount of horizontal solar radiation outside the atmosphere for 48 frames per day, going back one year prior to the prediction execution date. Next, the ratio calculation unit 121A' calculates the amount of electricity generated by the solar power generation equipment 200 for the past year using the calculated amount of horizontal solar radiation outside the atmosphere for the past year, calculated according to, for example, the calculation formulas (1) to (4) described above, and the installation orientation angle and tilt angle of the solar power generation panels 210. The installation orientation angle is the angle between a given orientation and the orientation in which the solar power generation panels 210 are installed, with the orientation being the base point. The tilt angle is the angle between the horizontal plane and the inclined surface of the solar power generation panels 210, with the horizontal plane being the base point. Once the ratio calculation unit 121A' has finished calculating the amount of power generated for the past year, it calculates the second ratio R of the amount of power generated by the solar power generation equipment 200. pv1 ~R pvx E is calculated. For example, the amount of power generated at the predicted time t and at t+1, one time frame after t, is calculated as E. pv (t),E pv Let (t+Δt) be the second ratio R pv1 ~R pvx is, E pv (t+Δt) / E pv (t) is calculated. The second ratio R for this year pv1 ~R pvx This is stored in the ratio DB133 within the storage device 130. Also, the second ratio R for one year pv1 ~R pvx By calculating this, the second ratio R for any time of year can be determined. pv1 ~R pvx It also becomes possible to use [this method].

[0053] The ratio calculation units 121A and 121A' are both functions implemented by the calculation processing of the arithmetic processing unit 121, but they are functions that can be implemented as alternatives. In other words, either the function of ratio calculation unit 121A' is implemented instead of ratio calculation unit 121A, or the function of ratio calculation unit 121A is implemented instead of ratio calculation unit 121A'. In this embodiment, for example, the function of ratio calculation unit 121A is implemented. However, it is also possible to have both ratio calculation units 121A' and ratio calculation unit 121A to calculate the first ratio and the second ratio, and then use their average.

[0054] Figure 7 shows an example of the maximum daily value of horizontal solar radiation outside the atmosphere for the past year (past fiscal year: April 1st to March 31st of the following year), calculated by the ratio calculation unit 121A. Figure 8 shows an example of the maximum daily value of power generated by the solar power generation equipment 200 for the past year, calculated by the ratio calculation unit 121A'. Note that the power generation shown in Figure 8 has been standardized so that the annual maximum value is equivalent to that in Figure 7.

[0055] Figure 9 shows an example of the waveform of the daily ratio of horizontal solar radiation outside the atmosphere (solid line) and the waveform of the daily ratio of power generation (dashed line). In Figure 9, the horizontal axis shows the number of frames per day (48), and the vertical axis shows the ratio. These ratio waveforms are shown as an example of the waveform for April 1st in the year shown in Figures 7 and 8. In the waveform of the ratio of horizontal solar radiation outside the atmosphere shown by the solid line in Figure 9, periods T1 and T5 are E sr (t) is the period before sunrise and after sunset, respectively, E sr Since (t) is 0, the ratio cannot be calculated in principle, but the ratio for this period is set to 1.0. Also, the period T2 following period T1 is E sr Since (t) is the period immediately after sunrise, the ratio for this period can actually be very large, but according to empirical rules, the maximum value is set to, for example, 1.8. Also, in period T3 following period T2, the ratio decreases in the morning from 1.8 to 1.0 in a nonlinear curve, and in the afternoon from 1.0 to 0 in a nonlinear curve. Also, period T4 between period T3 and period T5 is E sr(t) is the period just before sunset, and E sr (t + Δt) becomes 0, so the ratio also becomes 0. As described above, the ratio during the period T5 after sunset is 1.0. On the other hand, the periods T1’ to T5’ in the waveform of the ratio of the power generation amount shown by the broken line in FIG. 9 correspond to the periods T1 to T5 in the waveform of the ratio of the extraterrestrial horizontal solar irradiance shown by the solid line in FIG. 9. Since the data of the extraterrestrial horizontal solar irradiance is used to calculate the power generation amount, the waveform of the ratio of the power generation amount changes substantially in the same way as the waveform of the ratio of the extraterrestrial horizontal solar irradiance.

[0056] As a pre - processing before the solar power generation amount prediction system 100 predicts the power generation amount of the solar power generation facility 200 at a short time ahead from the prediction time on the prediction execution date, the frame determination unit 121B determines at which frame position around sunrise the t - th frame at the prediction time is. Specifically, the frame determination unit 121B has information indicating an r - th frame (the second frame) including the sunrise time, an r + a - th frame (from the second frame to the third frame a predetermined number of frames: the frame at which a sustainable model using the a - th frame added as a margin from the r - th frame can be used) required for the power generation amount of the solar power generation facility 200 to increase from the power generation amount at the sunrise time to a predetermined power generation amount, and an s - th frame indicating how many frames ahead from the t - th frame the power generation amount prediction is to be performed. For example, the following determination is made using accurate time information obtained from GPS satellites. Note that a and s are arbitrarily set parameters.

[0057] The frame determination unit 121B determines whether the t - th frame at the prediction time for predicting the power generation amount of the solar power generation facility 200 is a frame before (t < r + a) or a frame after (t ≧ r + a) the r + a - th frame (the third frame) after a frames have elapsed from the r - th frame including the sunrise time.

[0058] If the frame determination unit 121B determines that frame t, the prediction time for forecasting the power generation amount of the solar power generation equipment 200, is earlier than frame r+a, then either frame t is before sunrise and the power generation amount of the solar power generation equipment 200 is zero, or frame t is immediately after sunrise and the power generation amount of the solar power generation equipment 200 has not yet increased to a predetermined amount. When using a sustained model that predicts the power generation amount at the target time assuming that the current power generation amount will remain unchanged for a while, in the former case, even if the solar power generation equipment 200 has started generating power in frame t+1, the predicted value will be 0. In the latter case, the power generation amount of the solar power generation equipment 200 at frame t is unstable, and the predicted value at frame t+1 will also be inaccurate. For this reason, the past data search unit 121C accesses the past DB 131 and identifies past dates with weather equivalent to (same or similar to) the weather at the target time for forecasting the forecast execution date obtained from the Japan Meteorological Agency DB 400 as weather equivalent dates.

[0059] Figures 10 and 11 show an example of the search period for which the historical data search unit 121C searches for weather-equivalent days. The search period of the historical data search unit 121C is, for example, the 14 days prior to the day before the forecast execution date. Furthermore, if the historical DB 131 stores weather information for multiple years in the past, the search period may be the day one year prior to the forecast execution date and the 14 days before and after that day, the day two years prior to the forecast execution date and the 14 days before and after that day, the day three years prior to the forecast execution date and the 14 days before and after that day, etc. Since weather information is likely to be similar depending on the season, when searching for similar past weather information, it is effective to search around the month and day of the forecast execution date, such as the 14 days from the day before the forecast execution date or the 14 days before and after one year prior to the forecast execution date. However, this search period is just an example, and if there is a period during which past weather information similar to the weather information on the forecast execution date can be searched, that period may be used as the search period.

[0060] In Figure 11, the historical data search unit 121C accesses the historical DB 131 and determines that the weather information for n frames ahead of the prediction execution date corresponds to the weather information for n frames 2 days prior (reference day 2), 8 days prior (reference day 8), and 13 days prior (reference day 13) from the prediction execution date, and identifies these reference days 2, 8, and 13 as weather equivalent days. Furthermore, the historical data search unit 121C reads the actual power generation values ​​of the solar power generation facility 200 for n frames on the identified weather equivalent days from the historical DB 131. Note that for each identified reference day, PVact represents the actual power generation values ​​of the solar power generation facility 200 for the reference day for which similar historical weather information corresponds, Weather represents similar historical weather information, ssr represents the solar radiation in similar historical weather information, te represents the temperature in similar historical weather information, and wi represents the wind speed in similar historical weather information. Furthermore, similar past weather information (Weather) is information that associates similar past solar radiation (ssr), temperature (te), and wind speed (wi) for the reference date to which this weather information pertains.

[0061] The power generation forecasting unit 121D calculates the predicted power generation value for the period (forecast period) from t+1 (one frame ahead of the forecast time) to t+s (a short time ahead) depending on whether the number of weather-corresponding days identified by the past data search unit 121C is N or more (a predetermined number) or less than N. N is set to the minimum number necessary for the power generation forecasting unit 121D to calculate a highly accurate power generation prediction value from actual power generation values ​​for weather-corresponding days using fuzzy inference.

[0062] For example, if the prediction time frame t (the first frame) for predicting the power generation amount of the solar power generation facility 200 is earlier than frame r+a (the third frame), and the number of weather-corresponding days identified by the past data search unit 121C is N or more, the power generation prediction unit 121D uses fuzzy inference as one of the methods A to weight the actual power generation amount of the solar power generation facility 200 on weather-corresponding days based on the difference (solar radiation, temperature, wind speed) between the weather information on the prediction execution day and the weather information on weather-corresponding days, and calculates the predicted power generation amount of the solar power generation facility 200 from frame t+1 onwards. The following will explain this in detail with reference to Figures 12 to 16.

[0063] Figure 12 shows the power generation corrected value PVact(i), which is the actual power generation value weighted based on temperature. T This diagram shows the algorithm for calculating the power generation performance correction value PVact. T This is calculated using equation (5) shown in Figure 12.

[0064] Here, PVact(i) represents the actual power generation value on the reference day, which is a weather-equivalent day, and Tdist i This shows the absolute difference between the temperature information for the forecast execution date obtained from the Japan Meteorological Agency DB400 and the temperature information for the reference date obtained from past DB131. max Tdist i It shows the maximum value among the absolute values ​​shown by N in This indicates the number of reference days identified by the historical data search unit 121C.

[0065] And then, N in For each PVact(i), an algorithm is used to determine the grade using a membership function such that the closer it is to the temperature on the prediction execution day, the greater the weight. Based on this, the power generation prediction unit 121D calculates the power generation actual correction value PVact T Calculate.

[0066] Figure 13 shows the power generation corrected value PVact, which is the power generation actual value PVact(i) weighted based on wind speed. WThis diagram shows the algorithm for calculating the power generation performance correction value PVact. W This is calculated using equation (6) shown in Figure 13.

[0067] Here, PVact(i) represents the actual power generation value on the reference day, which is a weather-equivalent day, and Wdist i This shows the absolute difference between the wind speed information for the forecast execution date obtained from the Japan Meteorological Agency DB400 and the wind speed information for the reference date obtained from past DB131, and is displayed as Wdist. max is, Wdist i It shows the maximum value among the absolute values ​​shown by N in This indicates the number of reference days identified by the historical data search unit 121C.

[0068] And then, N in For each PVact(i), an algorithm is used to determine the grade using a membership function such that the closer it is to the wind speed on the prediction execution day, the greater the weight. Based on this, the power generation prediction unit 121D calculates the power generation actual correction value PVact W Calculate.

[0069] Figure 14 shows the power generation corrected value PVact, which is the power generation actual value PVact(i) weighted based on solar radiation. SR This diagram shows the algorithm for calculating the power generation performance correction value PVact. SR This is calculated using equation (7) shown in Figure 14.

[0070] Here, PVact(i) represents the actual power generation value on the reference day, which is a weather-equivalent day, and SRdist i This shows the absolute difference between the solar radiation information for the forecast execution date obtained from the Japan Meteorological Agency DB400 and the solar radiation information for the reference date obtained from past DB131, and SRdist max , SRdist i It shows the maximum value among the absolute values ​​shown by N in This indicates the number of reference days identified by the historical data search unit 121C.

[0071] And then, N inFor each PVact(i), an algorithm is used to determine the grade using a membership function such that the closer it is to the solar radiation on the prediction execution day, the greater the weight. Based on this, the power generation prediction unit 121D calculates the power generation actual correction value PVact. SR Calculate.

[0072] Figure 15 shows the three power generation performance correction values ​​PVact T PVact W PVact SR From the predicted power generation value PVact fore This diagram shows the algorithm for calculating the predicted power generation value PVact. fore This is calculated by formula (8) shown in Figure 15.

[0073] Specifically, the predicted power generation value PVact fore This is the PVact, which is a correction value for the actual amount of power generated. T PVact W PVact SR The values ​​are calculated using coefficients a, b, and c, which further weight each of these values. Note that coefficients a, b, and c are parameters that can be set arbitrarily. For example, one could conduct prediction experiments to calculate predicted power generation values ​​by changing coefficients a, b, and c, and then select and set the combination of coefficients a, b, and c that yields the highest prediction accuracy.

[0074] The predicted power generation value PVact in this embodiment fore This is the PVact, which is a correction value for the actual amount of power generated. T PVact W PVact SR It is calculated by weighting the power generation prediction value PVact fore This is the PVact, which is a correction value for the actual amount of power generated. T PVact W PVact SR One or three of the following power generation performance correction values ​​PVact T PVact W PVact SR It may also be a value calculated by weighting any two of the combinations of these.

[0075] FIG. 16 shows another example of the membership function used when calculating the power generation amount actual correction value PVact T PVact W PVact SR is a diagram showing another example of the membership function used when calculating

[0076] As shown in FIG. 16, by determining the grade using a non-linear membership function, the closer the temperature, wind speed, and solar radiation amount on the prediction execution date are, the greater the weighted grade can be determined. Therefore, when calculating the three power generation amount actual correction values PVact T PVact W PVact SR it may be possible to use the non-linear membership function shown in FIG. 16 instead of the linear membership function shown in FIGS. 12 to 14.

[0077] Next, when the t-th frame (the first frame) at the prediction time for predicting the power generation amount of the photovoltaic power generation facility 200 is before the (r + a)-th frame (the third frame), and the number of weather equivalent days specified by the past data search unit 121C is less than N (less than a predetermined number), the power generation amount prediction unit 121D may use method A, but calculates the power generation amount prediction value PV fore using the other method B described below to calculate the power generation amount prediction value with high accuracy. In the other method B, the past data search unit 121C specifies only N past weather equivalent days in order of the priority of a plurality of weathers previously associated with the weather of the n-th frame on the prediction execution date from the past DB131 during the search period shown in FIGS. 10, 11, etc. according to the table data stored in the search priority DB134, reads out the power generation amount actual value PVact(i) on the N weather equivalent days, and further, the power generation amount prediction unit 121D multiplies the power generation amount actual value PVact(i) read from the past data search unit 121C by the magnification factor Xi set for each of the plurality of weathers having priorities to calculate the power generation amount prediction value PV fore for each frame after the (t + 1)-th frame. This will be specifically described below with reference to FIGS. 6 and 17.

[0078] Figure 6 is a table showing the relationship between the priority of multiple weather conditions to be explored in the past and the multiplier Xi in the other method B, and Figure 17 shows the predicted power generation value PV. fore This figure shows the algorithm for calculating [the value]. Note that the table data shown in Figure 6 is pre-stored in the search priority DB 134 within the storage device 130.

[0079] The table data shown in Figure 6 illustrates the relationship between the weather (e.g., sunny, rainy, cloudy, snowy) for n frames ahead in the forecast time, obtained from the Japan Meteorological Agency DB400, multiple past weather conditions (e.g., sunny, rainy, cloudy, snowy) given priority to this weather condition, and the magnification Xi associated with these multiple past weather conditions.

[0080] If the weather in n frames ahead of the predicted time is sunny, then past weather conditions are given priority in the order of sunny, cloudy, rainy, and snowy. A magnification Xi is associated with each of the past weather conditions so that the amount of solar radiation obtained is equivalent to that of sunny weather ahead of the predicted time. For example, if the past weather was sunny, the magnification Xi=1; if the past weather was cloudy, the magnification Xi=2; if the past weather was rainy, the magnification Xi=3; and if the past weather was snowy, the magnification Xi=5. According to this relationship, if the weather in n frames ahead of the predicted time is sunny, the past data search unit 121C searches for past weather conditions in the order of sunny, cloudy, rainy, and snowy until the number of searches reaches N during the search period shown in Figures 10 and 11.

[0081] Furthermore, if the weather in n frames ahead of the predicted time is rain, then multiple past weather conditions are given priority in the order of rain, cloudy, sunny, and snow. A magnification Xi is associated with each of the multiple past weather conditions so that the amount of solar radiation obtained is equivalent to the amount of solar radiation due to rain at the predicted time. For example, if the past weather was rain, the magnification Xi=1; if the past weather was cloudy, the magnification Xi=0.6; if the past weather was sunny, the magnification Xi=0.2; and if the past weather was snow, the magnification Xi=1.1. According to this relationship, if the weather in n frames ahead of the predicted time is rain, the past data search unit 121C searches for past weather conditions in the order of rain, cloudy, sunny, and snow during the search period shown in Figures 10 and 11, etc., until the number of searches reaches N.

[0082] Furthermore, if the weather in n frames ahead of the predicted time is cloudy, then multiple past weather conditions are given priority in the order of cloudy, rainy, sunny, and snowy. A magnification Xi is associated with each of the multiple past weather conditions so that the amount of solar radiation obtained is equivalent to the amount of solar radiation due to cloudy weather ahead of the predicted time. For example, if the past weather was cloudy, the magnification Xi=1; if the past weather was rainy, the magnification Xi=1.5; if the past weather was sunny, the magnification Xi=0.5; and if the past weather was snowy, the magnification Xi=1.5. According to this relationship, if the weather in n frames ahead of the predicted time is cloudy, the past data search unit 121C searches for past weather conditions in the order of cloudy, rainy, sunny, and snowy until the number of searches reaches N during the search period shown in Figures 10 and 11.

[0083] Also, when the weather in the n frames ahead of the prediction time is snow, the past plurality of weathers are given priorities, for example, in the order of snow, rain, cloudy, and sunny. For the past plurality of weathers, a magnification factor Xi that can obtain a solar radiation amount equivalent to that of the solar radiation amount due to snow ahead of the prediction time is associated. For example, when the past weather is snow, the magnification factor Xi = 1; when the past weather is rain, the magnification factor Xi = 0.9; when the past weather is cloudy, the magnification factor Xi = 0.5; when the past weather is sunny, the magnification factor Xi = 0.2. According to this relationship, when the weather in the n frames ahead of the prediction time is snow, the past data search unit 121C searches for the past weather in the order of snow, rain, cloudy, and sunny until the search count reaches N during the search period shown in FIGS. 10, 11, etc.

[0084] Note that the priority of the past weather and the magnification factor Xi are arbitrarily set for each region where the solar power generation facility 200 is installed.

[0085] Then, when the search for the weather equivalent day by the past data search unit 121C ends, the power generation amount prediction unit 121D calculates the predicted power generation value PV according to the formula (9) shown in FIG. 17. Note that N fore is the total number N of the weather equivalent days specified by the past data search unit 121C. days

[0086] ​The power generation forecasting unit 121D calculates the predicted power generation value of the solar power generation equipment 200 for each frame within the forecast period, from frame t+1 (one frame beyond the current frame t) to frame t+s. In the process of the power generation forecasting unit 121D calculating the predicted power generation value of the solar power generation equipment 200 frame by frame, there is a possibility that one of the s frames between frames t+1 and t+s will be later than frame r+a. In other words, if one of the s frames between frames t+1 and t+s (n frames) is later than frame r+a, then for subsequent frames, it becomes possible to use a sustained model to calculate a highly accurate predicted power generation value of the solar power generation equipment 200. Therefore, if the frame determination unit 121B determines that frame t is earlier than frame r+a, the next determination process is to determine whether any frame between frame t+1 and frame t+s is later than frame r+a, which makes it possible to use the sustained model to predict the amount of power generated by the solar power generation equipment 200.

[0087] If the frame determination unit 121B determines that frame n, which is the target time (forecast time) between frames t+1 and t+s, is still before frame r+a, the power generation prediction unit 121D uses the calculated power generation prediction value shown in equation (8) or equation (9). Alternatively, instead of using the power generation prediction value calculated by equation (8) or equation (9), the power generation prediction value for frame n calculated by an engineering model that takes the latest forecast values ​​such as solar radiation, temperature, and wind speed from the Meso Scale Model (MSM), a meteorological model introduced by the Japan Meteorological Agency as data for creating disaster prevention meteorological information, may be used. Alternatively, the power generation prediction value for frame n calculated by a method described in published patent applications such as JP 2020-123198, JP 2020-123199, JP 2020-123200, and JP 2022-23410, for which the applicant is the applicant, may be used.

[0088] On the other hand, if the frame determination unit 121B determines that the nth frame, which is the target time (forecast time) between frames t+1 and t+s, is later than frame r+a, the power generation prediction unit 121D switches to a method of calculating the predicted power generation value of the solar power generation equipment 200 using a sustained model. For example, the power generation prediction unit 121D reads the ratio corresponding to frame n from the ratio DB 133 and calculates the predicted power generation value by multiplying the predicted power generation value calculated by equation (8) or equation (9) for the n-1st frame (the frame before) by this ratio. Note that instead of using the predicted power generation value calculated by equation (8) or equation (9) as the target of multiplication by the ratio, the predicted power generation value for frame n-1 calculated by an engineering model that utilizes the forecast values ​​of the mesoscale numerical weather prediction model MSM may be used, or the predicted power generation value for frame n-1 calculated by the method described in the above-mentioned published gazette, for which the present applicant is the applicant, may be used.

[0089] Furthermore, if the frame determination unit 121B determines that frame t (the first frame) at the prediction time (the current time) is later than frame r+a (the third frame), the power generation prediction unit 121D will use a sustained model to calculate the predicted power generation value of the solar power generation equipment 200. Specifically, the power generation prediction unit 121D obtains the actual power generation value that the solar power generation equipment 200 actually generated in frame t, reads the ratio corresponding to frame t+1 from the ratio DB 133, and calculates the predicted power generation value for frame t+1 by multiplying the obtained actual power generation value for frame t by this ratio.

[0090] Figure 18 shows the algorithm by which the power generation prediction unit 121D calculates a power generation prediction value based on the predicted value for n-1 frames or the actual power generation value for t frames generated by the solar power generation equipment 200. This power generation prediction value is calculated by equation (10) shown in Figure 18. Note that in Figure 18, E sr (t+Δt) / E sr(t) represents the first ratio obtained by dividing the amount of horizontal solar radiation outside the atmosphere in each frame by the amount of horizontal solar radiation outside the atmosphere in the previous frame, E(t) represents the actual or predicted amount of power generation at the prediction time (current time), and E(t+Δt) represents the predicted amount of power generation one frame ahead from the prediction time. As shown in equation (10), the predicted amount of power generation one frame ahead can be calculated by multiplying the actual or predicted amount of power generation at the current time by the first ratio of horizontal solar radiation outside the atmosphere. When using the ratio calculation unit 121A' instead of the ratio calculation unit 121A, the first ratio E shown in equation (10) sr (t+Δt) / E sr (t) is the second ratio E pv (t+Δt) / E pv This will be replaced by (t).

[0091] Figure 19 is a flowchart showing an example of how the solar power generation forecasting system 100 works when calculating the predicted power generation value of the solar power generation equipment 200.

[0092] First, the ratio calculation unit 121A calculates the amount of horizontal solar radiation outside the atmosphere for one year prior to the forecast execution date (April 1st to March 31st of the following year) at the location where the solar power generation equipment 200 is installed. Once the ratio calculation unit 121A has completed calculating the amount of horizontal solar radiation outside the atmosphere for one year prior to the forecast execution date, it calculates the first ratio E of the horizontal solar radiation outside the atmosphere. sr (t+Δt) / E sr (t) is calculated. This first ratio for one year of past years is stored in the ratio DB133 in the storage device 130. (S1010)

[0093] Each time slot is a 30-minute unit, and there are 48 time slots in a day. Therefore, the calculation processing unit 121 executes a second loop process including steps S1050, S1060, and S1070, as described below, and a first loop process including this second loop process and steps S1020, S1030, S1040, and S1080, for t time slots (1 ≤ t ≤ 48) at the time of prediction.

[0094] The frame determination unit 121B determines whether frame t, representing the prediction time, is earlier or later than frame r+a, which represents the increase in the amount of power generated by the solar power generation equipment 200 from the amount of power generated at sunrise to a predetermined amount. (S1020)

[0095] If the frame determination unit 121B determines that frame t at the prediction time is earlier than frame r+a (S1020: YES), then, since it is a time period in which an accurate power generation prediction cannot be calculated even if a sustained model is used to calculate the predicted power generation amount of the solar power generation equipment 200, the past data search unit 121C accesses the past DB 131 and identifies weather days corresponding to the weather information (weather, solar radiation, temperature, wind speed) for frames t+1 to t+s of the prediction execution date obtained from the Japan Meteorological Agency DB 400 during the search period shown in Figures 10 and 11, etc. (S1030)

[0096] The power generation forecasting unit 121D calculates the predicted power generation values ​​for the solar power generation facility 200 from frame t+1 to frame t+s, which is one frame beyond frame t of the forecast time, using different methods A and B depending on whether the number of weather-corresponding days identified by the past data search unit 121C is N or more, or less than N. (S1040)

[0097] Here, a specific example of step S1040 described above will be explained with reference to Figure 20.

[0098] The power generation forecasting unit 121D determines whether the number of weather-corresponding days identified by the past data search unit 121C is N or more, or less than N. (S2010)

[0099] For example, if the prediction time frame t for predicting the power generation amount of the solar power generation facility 200 is earlier than frame r+a, and furthermore, the number of weather-corresponding days identified by the past data search unit 121C is N or more (S2010:YES), the power generation prediction unit 121D reads the actual power generation value of the solar power generation facility 200 on the reference day, which is a weather-corresponding day, from the past DB 131. Using one of the methods A, it weights this actual power generation value based on the difference (solar radiation, temperature, wind speed) between the weather information on the prediction execution day and the weather information on the weather-corresponding days using fuzzy inference (Equations (5) to (8)), and then calculates the predicted power generation value PVact of the solar power generation facility 200 from frame t+1 onwards. fore Calculate (S2020).

[0100] On the other hand, if the time frame t at which the power generation amount of the solar power generation equipment 200 is predicted is earlier than time frame r+a, and the number of weather-corresponding days identified by the past data search unit 121C is less than N (S2010:NO), the past data search unit 121C, while referring to the table data in Figure 6 stored in the search priority DB 134, identifies only N past weather-corresponding days from the past DB 131 in the order of the priority of multiple weather conditions pre-associated with the weather of n time frames ahead of the prediction time of the prediction execution date, during the search period shown in Figures 10, 11, etc., and reads out the actual power generation amount PVact(i) for the N weather-corresponding days. (S2030)

[0101] Next, the power generation prediction unit 121D, referring to the table data in Figure 6, uses equation (9) to multiply the actual power generation value PVact(i) read by the past data search unit 121C by a multiplier Xi set for each of the multiple weather conditions with priority, to obtain the predicted power generation value PV from frame t+1 onwards. fore Calculate (S2040).

[0102] Returning to Figure 19, in step S1040 above, the predicted power generation values ​​for the solar power generation equipment 200 from t+1 onwards are calculated using equation (8) or equation (9) according to the number of weather-corresponding days identified by the past data search unit 121C, and the second loop processing described below is performed.

[0103] The frame determination unit 121B determines whether frame n, which is the prediction target time between frame t+1 and frame t+s for which the power generation prediction unit 121D calculates the power generation prediction value, is a frame before frame r+a or a frame after frame r+a. (S1050)

[0104] If n frames of the prediction target time between frames t+1 and t+s are earlier than frame r+a (S1050: YES), the power generation prediction unit 121D outputs as a calculation result a power generation prediction value calculated using equation (8) or equation (9), or a power generation prediction value for n frames calculated by an engineering model that utilizes the forecast values ​​of the mesoscale numerical weather prediction model MSM, or a power generation prediction value for n frames calculated by the method described in the above-mentioned published gazette for which the present applicant is the applicant. (S1060)

[0105] On the other hand, if n frames, which are the prediction target time between frames t+1 and t+s, are later than frames r+a (S1050:NO), the power generation prediction unit 121D switches to a method of calculating the predicted power generation value of the solar power generation equipment 200 using a sustained model. For example, the power generation prediction unit 121D reads a first ratio corresponding to n frames from the ratio DB 133 and calculates the predicted power generation value by multiplying the predicted power generation value calculated by equation (8) or equation (9) for the previous frame (n-1 frames) by this first ratio, or by multiplying the predicted power generation value for n frames calculated by an engineering model that uses the forecast values ​​of the mesoscale numerical weather prediction model MSM by this ratio, or by multiplying the predicted power generation value for n frames calculated by multiplying the predicted power generation value for n frames calculated by the method described in the above-mentioned published gazette, of which the present applicant is the applicant, by this ratio and outputs it as the calculation result. (S1070)

[0106] Returning to step S1020 above, if the frame determination unit 121B determines that frame t at the prediction time is later than frame r+a (S1020: NO), the power generation prediction unit 121D obtains the actual power generation value of the solar power generation equipment 200 at frame t from the past DB 131, reads the first ratio corresponding to frame t+1 from the ratio DB 133, and calculates the predicted power generation value for frame t+1 by multiplying the obtained actual power generation value for frame t by this ratio. (S1080)

[0107] The first loop process, including the second loop process described above, is executed up to frame t+s, which predicts the amount of electricity generated by the solar power generation equipment 200.

[0108] Figure 21 is a flowchart illustrating another example of the operation of the solar power generation forecasting system 100 when calculating the predicted power generation value of the solar power generation facility 200. Steps S3010, S3020, and S3040 in Figure 21 are the same as steps S1010, S1020, and S1080 in Figure 19, so their explanation is omitted. Furthermore, if the frame determination unit 121B determines that frame t at the time of prediction is earlier than frame r+a (S3020: YES), the power generation forecasting unit 121D outputs as the calculation result the predicted power generation value for n frames (t+1≦n≦t+s) calculated by an engineering model that utilizes the forecast values ​​of the mesoscale numerical weather prediction model MSM, or the predicted power generation value for n frames calculated by the method described in the above-mentioned published gazette, of which the present applicant is the applicant. (S3030)

[0109] The loop processing, including the steps S3020, S3030, and S3040 described above, is executed up to frame t+s, where the amount of power generated by the solar power generation equipment 200 is predicted.

[0110] As described above, the photovoltaic power generation prediction system 100 that calculates the predicted power generation amount of the photovoltaic power generation facility 200 includes a storage device 130 that stores the past actual power generation amount values of the photovoltaic power generation facility 200 and the past weather information including weather and meteorological data in association with the date and time information, and a period to be predicted (t + 1 to t + s) that starts from the next period after the t-th period including the current time point (prediction time point) in units of periods (komas) formed by dividing one day into fixed time intervals. The power generation prediction unit 121D calculates the predicted power generation amount value of the photovoltaic power generation facility 200 for each period within the period to be predicted. When the t-th period is a period after a predetermined number (r + a) of periods after the r-th period including the sunrise time (t ≧ r + a), the power generation prediction unit 121D performs a first process of calculating the predicted power generation amount value for each period within the period to be predicted using the actual power generation amount value or the predicted power generation amount value in the immediately preceding period. When the t-th period is a period before the a-th period (t < r + a), the power generation prediction unit 121D specifies the past date with the same weather as the weather during the period to be predicted as the weather-equivalent date, and performs a second process of calculating the predicted power generation amount value for each period within the period to be predicted using the actual power generation amount values for each period in the same time zone as the period to be predicted on the weather-equivalent date.

[0111] According to the photovoltaic power generation prediction system 100, for example, it is possible to accurately calculate the predicted power generation amount value of the photovoltaic power generation facility 200 a few minutes to a few hours in the short term from the prediction time point including the dawn time zone.

[0112] In addition, in the photovoltaic power generation prediction system 100, when the number of past dates specified as the weather-equivalent date is N or more in the second process, the power generation prediction unit 121D performs a predetermined weighting using fuzzy inference on the actual power generation amount values for each period in the same time zone on each weather-equivalent date according to the degree of similarity between the meteorological data (solar radiation amount, temperature, wind speed, etc.) during the period to be predicted and the meteorological data (solar radiation amount, temperature, wind speed, etc.) on each weather-equivalent date, and then calculates the predicted power generation amount value for each period within the period to be predicted.

[0113] According to the solar power generation forecasting system 100, it is possible to identify N or more past dates as weather-corresponding days, making it possible to accurately calculate the predicted power generation value of the solar power generation facility 200 in the short time frame from the forecasting point using fuzzy inference.

[0114] Furthermore, in the solar power generation forecasting system 100, if the number of past dates identified as weather equivalent days is less than N, the power generation forecasting unit 121D adds past dates with weather predetermined according to the type of weather in the forecast period as weather equivalent days until the number of weather equivalent days reaches N. The unit then multiplies the actual power generation value for each time slot in the same time period on each weather equivalent day by a multiplier Xi determined according to the degree of similarity between the weather in the forecast period and the weather on each weather equivalent day, and calculates the predicted power generation value for each time slot within the forecast period.

[0115] According to the solar power generation forecasting system 100, even if it is not possible to identify N or more past dates as weather-corresponding days, it is possible to accurately calculate the predicted power generation value of the solar power generation facility 200 in the short time frame from the forecasting point.

[0116] Furthermore, in the solar power generation prediction system 100, the power generation prediction unit 121D, in the second process, if at least the last frame of the prediction period is frame r+a or later, performs the second process for frames before frame r+a within the prediction period, and performs the first process for frames r+a or later within the prediction period.

[0117] According to the solar power generation forecasting system, it is possible to accurately calculate the predicted power generation value of 200 solar power generation facilities a short time ahead from the forecasting point, regardless of whether the final time frame of the forecasting period falls before or after the third time frame.

[0118] Furthermore, the solar power generation forecasting system 100 includes a ratio calculation unit 121A that calculates a first ratio, which is the rate of change of the external horizontal solar radiation from the previous frame, for each frame by dividing the value of the external horizontal solar radiation at the installation site of the solar power generation equipment 200 for each frame by the value of the previous frame. The power generation forecasting unit 121D calculates the predicted power generation value for each frame within the forecast period in the first processing by multiplying the actual power generation value or predicted power generation value for the previous frame by the first ratio.

[0119] According to the solar power generation prediction system 100 of the present invention, it is possible to accurately calculate the predicted power generation value of a solar power generation facility 200 a short time from the prediction point in time using a first ratio.

[0120] Furthermore, the solar power generation forecasting system 100 includes a ratio calculation unit 121A' that calculates a second ratio, which is the rate of change of the calculated value of the solar power generation equipment 200 from the previous frame, for each frame by dividing the calculated value of the solar power generation equipment 200, which is calculated for each frame using a predetermined calculation formula, by the calculated value in the previous frame. The power generation forecasting unit 121D calculates the predicted power generation value for each frame within the forecasting period in the first processing by multiplying the actual power generation value or predicted power generation value in the previous frame by the second ratio.

[0121] According to the solar power generation prediction system 100 of the present invention, it is possible to accurately calculate the predicted power generation value of the solar power generation facility 200 a short time from the prediction point in time using a second ratio.

[0122] This embodiment is provided to facilitate understanding of the present invention and is not intended to limit its interpretation. The present invention may be modified or improved without departing from its spirit, and equivalents thereof are also included. [Explanation of Symbols]

[0123] 1. Power Systems 100 Solar Power Generation Prediction System 110 Input Device 120 Control device 121 Arithmetic Processing Unit 121A,121A' Ratio calculation section 121B Frame determination unit 121C Historical Data Search Unit 121D Power generation forecasting unit 122 memory 130 Storage device 131 Past DB 132 Equipment DB 133 Ratio DB 134 Search priority DB 140 Communication equipment 200 Solar power generation facilities 210 Solar panels 220 Power Conditioner 300 weather observation instruments 400 Japan Meteorological Agency Database 500 Communication Networks 600 Power system

Claims

1. A solar power generation forecasting system that calculates predicted power generation values ​​for solar power generation facilities, A storage device that stores historical power generation data of the aforementioned solar power generation facility and historical weather information including weather and meteorological data, in association with date and time information. A prediction device that calculates predicted power generation values ​​for the solar power generation equipment in each time slot within a prediction period, starting from the time slot following the first time slot which includes the current time, using time slots which are formed by dividing the day into fixed time intervals; A first ratio calculation device calculates a first ratio, which is the rate of change of the horizontal solar radiation outside the atmosphere from the previous frame, for each frame by dividing the value of the horizontal solar radiation outside the atmosphere at the installation site of the solar power generation equipment in each frame by the value of the previous frame. Equipped with, The prediction device is If the first frame is a predetermined number of frames from the second frame which includes the time of sunrise, then a first process is performed to calculate the predicted power generation value for each frame within the prediction period using the actual power generation value or predicted power generation value for the immediately preceding frame. If the first frame is a frame earlier than the third frame, a past date with weather corresponding to the weather of the forecast period is identified as the weather equivalent day, and a second process is performed to calculate the predicted power generation value for each frame within the forecast period using the actual power generation value for each frame in the same time period as the forecast period on the weather equivalent day. The prediction device is In the first process, the predicted power generation value for each frame within the prediction period is calculated by multiplying the actual power generation value or predicted power generation value for the immediately preceding frame by the first ratio. In the second process described above, if the number of past dates identified as corresponding weather days exceeds a predetermined number, a predetermined weighting is applied using fuzzy inference to the actual power generation values ​​for each time slot within the same time period on each corresponding weather day, according to the degree of similarity between the weather data for the forecast period and the weather data for each corresponding weather day, and then the predicted power generation values ​​for each time slot within the forecast period are calculated. In the second process described above, if the number of past dates identified as weather equivalent days is less than a predetermined number, past dates that had weather predetermined according to the type of weather during the forecast period are added as weather equivalent days until the number of weather equivalent days reaches the predetermined number. The predicted power generation value for each time slot within the forecast period is calculated by multiplying the actual power generation value for each time slot in the same time period on each weather-equivalent day by a multiplier determined according to the degree of similarity between the weather during the forecast period and the weather on each corresponding day. Solar power generation forecasting system.

2. A solar power generation prediction system according to claim 1, The prediction device is In the second process, if at least the last frame of the prediction period is the third frame or later, the second process is performed for frames that are before the third frame within the prediction period, and the first process is performed for frames that are the third frame or later within the prediction period. Solar power generation forecasting system.

3. A solar power generation prediction system according to claim 1, The system includes a second ratio calculation device that calculates a second ratio, which is the rate of change of the calculated value of the solar power generation equipment from the previous frame, for each frame by dividing the calculated value of the solar power generation equipment, which is calculated for each frame using a predetermined calculation formula, by the calculated value for the previous frame. In the first process, the prediction device calculates the predicted power generation value for each frame within the prediction period by multiplying the actual power generation value or predicted power generation value for the immediately preceding frame by the second ratio. The aforementioned predetermined calculation formulas are the following formulas (1) to (4): Tpa=Ta+(A / (B×V 0.8 +1)+2)×Ga-2...(1) (Tpa: Solar panel temperature (°C), Ta: Ambient temperature (°C), A: Coefficient (e.g., roof-mounted type "50"), B: Coefficient (e.g., roof-mounted type "0.38"), V: Wind speed (m / s), Ga: Solar radiation on inclined surface (kW / m²)) Kpt=1+αmax×(Tpa-25)...(2) (Kpt: Temperature correction coefficient, αmax: Maximum power temperature coefficient (1 / °C), Tpa: Solar panel temperature (°C)) System output coefficient = Kloss × Kpt ... (3) (Kloss: A fixed coefficient used universally depending on the type of equipment to take into account changes over time (dirt, deterioration), wiring resistance loss, inverter loss, etc.; Kpt: Temperature correction coefficient) Calculated value (Pw) = Solar radiation on inclined surface (Ga) × System output coefficient × Panel capacity ... (4) Solar power generation forecasting system.

4. A method for predicting solar power generation, which calculates a predicted value for the amount of power generated by a solar power generation facility, The first step involves storing the past power generation data of the aforementioned solar power generation facility and past weather information, including weather and meteorological data, in a storage device in association with date and time information. The second step involves using a prediction device to calculate the predicted power generation amount of the solar power generation equipment in each time slot within the prediction period, starting from the time slot following the first time slot which includes the current time, with time slots being the units formed by dividing the day into fixed time intervals. A third step involves calculating a first ratio, which is the rate of change of the horizontal solar radiation outside the atmosphere from the previous frame, for each frame by dividing the value of the horizontal solar radiation outside the atmosphere at the installation site of the solar power generation equipment by the value of the previous frame, using a first ratio calculation device. Includes, In the second step described above, If the first frame is a predetermined number of frames from the second frame which includes the time of sunrise, then a first process is performed to calculate the predicted power generation value for each frame within the prediction period using the actual power generation value or predicted power generation value for the immediately preceding frame. If the first frame is a frame earlier than the third frame, a past date with weather corresponding to the weather of the forecast period is identified as the weather equivalent day, and a second process is performed to calculate the predicted power generation value for each frame within the forecast period using the actual power generation value for each frame in the same time period as the forecast period on the weather equivalent day. In the first process, the predicted power generation value for each frame within the prediction period is calculated by multiplying the actual power generation value or predicted power generation value for the immediately preceding frame by the first ratio. In the second process described above, if the number of past dates identified as corresponding weather days exceeds a predetermined number, a predetermined weighting is applied using fuzzy inference to the actual power generation values ​​for each time slot within the same time period on each corresponding weather day, according to the degree of similarity between the weather data for the forecast period and the weather data for each corresponding weather day, and then the predicted power generation values ​​for each time slot within the forecast period are calculated. In the second process described above, if the number of past dates identified as weather equivalent days is less than a predetermined number, past dates that had weather predetermined according to the type of weather during the forecast period are added as weather equivalent days until the number of weather equivalent days reaches the predetermined number. The predicted power generation value for each time slot within the forecast period is calculated by multiplying the actual power generation value for each time slot in the same time period on each weather-equivalent day by a multiplier determined according to the degree of similarity between the weather during the forecast period and the weather on each corresponding day. Methods for predicting solar power generation.

Images