Power generation forecasting system

Abstract

We provide a power generation forecasting system that can improve the accuracy of power generation forecasts. [Solution] A power generation prediction system 100 that predicts the amount of power generated by a power generation unit 200 that can generate power using natural energy, comprising: a power generation prediction value acquisition unit that predicts the amount of power generated based on predetermined information and acquires the predicted amount of power generated as a power generation prediction value (step S11); a storage unit that stores the power generation prediction value and the actual power generation value (step S19); an output rate acquisition unit that acquires the ratio of the actual power generation value on a past output control implementation date to the power generation prediction value as the output rate (step S13); and a correction unit that performs a correction process to correct the power generation prediction value using the output rate if there is a possibility that output control will be implemented on the target date of the prediction (YES in step S12) (step S13).

Description

Technical Field

[0001] The present invention relates to a technology of a power generation amount prediction system that predicts the power generation amount by a power generation unit capable of generating power using natural energy.

Background Art

[0002] Conventionally, a power generation amount prediction system that predicts the power generation amount by a power generation unit capable of generating power using natural energy has been known. For example, it is as described in Patent Document 1.

[0003] Patent Document 1 discloses a method for predicting the power generation amount by a photovoltaic power generation device in a photovoltaic power generation business. By predicting the power generation amount in this way, the predicted value of the power generation amount can be utilized for transactions in the wholesale power market or the like.

[0004] Here, when the power generation amount by the photovoltaic power generation device becomes excessive with respect to the power demand in the entire power supply area, the power company (general power transmission and distribution business operator) may request the power generation operator to perform output control. When output control is requested, there is a problem that a deviation occurs between the predicted value and the actual value of the power generation amount, and the accuracy of the power generation amount prediction decreases.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] The present invention has been made in view of the above circumstances, and the problem to be solved is to provide a power generation amount prediction system capable of improving the accuracy of power generation amount prediction.

Means for Solving the Problems

[0007] The problems that this invention aims to solve are as described above, and the means for solving these problems will now be explained.

[0008] In other words, claim 1 is a power generation prediction system that predicts the amount of power generated by a power generation unit capable of generating power using natural energy, comprising: a power generation prediction value acquisition unit that predicts the amount of power generated based on predetermined information and acquires the predicted amount of power generated as a power generation prediction value; a storage unit that stores the power generation prediction value and the power generation actual value which is the actual value of the amount of power generated; an output rate acquisition unit that acquires the ratio of the power generation actual value on a past output control implementation date to the power generation prediction value on that implementation date as an output rate; and a correction unit that performs a correction process to correct the power generation prediction value using the output rate if there is a possibility that output control will be implemented on the target date of the prediction.

[0009] In claim 2, the correction unit performs the correction process based on calendar information.

[0010] In claim 3, the correction unit predicts the time periods in which the output control is performed based on the output rate for each predetermined time period, sets the predicted time periods as the predicted time periods, performs the correction process for the predicted time periods, and does not perform the correction process for time periods other than the predicted time periods.

[0011] In claim 4, the correction unit resets the forecast target time period based on the time period in which the confirmed value of the electricity market price is less than or equal to a predetermined price, after the electricity market price has been determined.

[0012] In claim 5, if the prediction target time period is expanded as a result of the reset, and the prediction target time period includes a time period in which the output rate is equal to or greater than a predetermined threshold, the correction unit performs the correction processing for that time period using the output rate of the time period closest to that time period in which the output rate is less than the predetermined threshold.

[0013] In claim 6, when an output control instruction including information on the target time period and output control amount for the output control is announced, the correction unit re-executes the correction process based on the target time period and output control amount specified in the output control instruction.

[0014] In claim 7, the predetermined information includes learning data including past power generation performance values, and the storage unit stores the uncorrected power generation prediction value initially acquired by the power generation prediction value acquisition unit as the learning data for the target time period in which the output control was performed, instead of the power generation performance values. [Effects of the Invention]

[0015] The present invention provides the following effects:

[0016] Claim 1 makes it possible to improve the accuracy of predicting the amount of power generated.

[0017] Claim 2 allows for further improvement in the accuracy of power generation prediction.

[0018] In claim 3, the correction process can be simplified.

[0019] Claim 4 allows for accurate prediction of the time period during which output control is to be performed.

[0020] Claim 5 makes it possible to improve the accuracy of power generation prediction values ​​in an expanded prediction target time period (one time period).

[0021] Claim 6 allows for further improvement in the accuracy of power generation prediction.

[0022] Claim 7 allows for further improvement in the accuracy of power generation prediction. [Brief explanation of the drawing]

[0023] [Figure 1] A block diagram showing the configuration of a power generation prediction system according to one embodiment of the present invention. [Figure 2] A flowchart showing predictive control for predicting power generation amount. [Figure 3] (a) A graph showing the transition of predicted power generation values. (b) A graph showing the transition of output rates. [Figure 4] (a) A graph showing the predicted power generation value after correction based on the output rate. (b) A graph showing the predicted power generation value after correction based on the determined value of the electricity market price. [Figure 5] A graph showing the predicted power generation value after correction based on the output control instruction.

Mode for Carrying Out the Invention

[0024] Hereinafter, a power generation amount prediction system 100 according to an embodiment of the present invention will be described.

[0025] The power generation amount prediction system 100 according to an embodiment of the present invention predicts the power generation amount by a power generation unit 200 that can generate power using natural energy. In the present embodiment, the power generation unit 200 is composed of a plurality of power generation devices that can generate power using sunlight. The power generation unit 200 is configured to be able to sell the generated power to an electric power company. The user of the power generation amount prediction system 100 is, for example, a power generation business operator. As shown in FIG. 1, the power generation amount prediction system 100 includes an acquisition unit 110 and a calculation unit 120.

[0026] The acquisition unit 110 acquires various information. The acquisition unit 110 can acquire information for predicting the power generation amount (hereinafter, sometimes simply referred to as "power generation amount") by the power generation unit 200 via a network such as the Internet. The information acquired by the acquisition unit 110 includes, for example, temperature, solar radiation amount, wind speed, power demand, fuel price, and the like.

[0027] The calculation unit 120 performs various calculations. Based on predetermined information, the calculation unit 120 can predict the amount of power generated on a target day. The predetermined information includes information such as temperature, solar radiation, wind speed, electricity demand, and fuel price, which are acquired by the acquisition unit 110. The predetermined information also includes training data for predicting the amount of power generated. The training data includes past data of predicted power generation values ​​(hereinafter referred to as "predicted power generation values") from the power generation unit 200, and past data of actual power generation values ​​(hereinafter referred to as "actual power generation values") from the power generation unit 200.

[0028] Thus, the calculation unit 120 obtains power generation prediction values ​​based on the aforementioned information. Among the information obtained by the acquisition unit 110, solar radiation, in particular, has a strong correlation with power generation. Therefore, in this embodiment, the calculation unit 120 predicts power generation using a prediction model (a machine learning model created using training data) that utilizes the correlation between solar radiation and power generation. The calculation unit 120 uses this prediction model to obtain power generation prediction values ​​for each unit time on the day to be predicted. The unit time is, for example, 30 minutes (1 frame). That is, the calculation unit 120 obtains power generation prediction values ​​for each of the 48 frames on the day to be predicted.

[0029] The calculation unit 120 is equipped with a storage unit capable of storing various types of data. Various types of information acquired via the acquisition unit 110 are stored in the storage unit. The calculation unit 120 (storage unit) is also electrically connected to the power conditioner (PCS) of the power generation unit 200, and the storage unit stores information such as the actual power generation values ​​from the power generation unit 200. In addition, the storage unit stores the calculation results (such as predicted power generation values) from the calculation unit 120.

[0030] With the power generation forecasting system 100 configured in this way, it is possible to obtain power generation forecast values ​​before bidding in the wholesale electricity market. The user can then use the obtained power generation forecast values ​​to predict the electricity market price and forecast the bidding curve (supply and demand curve) for bidding in the wholesale electricity market.

[0031] In cases where the amount of electricity generated exceeds the electricity demand across the entire power supply area of ​​a power company (general transmission and distribution operator), the power company may request power generators to implement output control. Here, "output control" means that power generators, at the direction of the power company, reduce (control) the output of their solar or wind power plants. Output control instructions are announced by the day before the day the output control is to be implemented.

[0032] When an output control instruction is given, the user must suppress the amount of power generated by the power generation unit 200 (more specifically, the output from the power conditioner (PCS) of the power generation unit 200) according to the content of the output control. As a result, the actual power generation value (more specifically, the actual output from the PCS of the power generation unit 200) will be less than the predicted power generation value, causing a discrepancy between the predicted and actual power generation values.

[0033] Therefore, in the power generation prediction system 100 according to this embodiment, the calculation unit 120 performs the prediction control shown in Figure 2 to predict the amount of power generated by the power generation unit 200. The method for predicting the amount of power generated by the power generation unit 200 will be explained below using the flowchart shown in Figure 2.

[0034] In the following, the day for which power generation is predicted will be referred to as "the current day," and the day before the predicted day will be referred to as "the previous day." The prediction control shown in Figure 2 is executed once a day before 10:00 a.m. on the day before, which is the deadline for bidding in the wholesale electricity market.

[0035] In step S11, the calculation unit 120 predicts the amount of power generated every 30 minutes (48 frames) on that day (the day to be predicted). Based on information such as the amount of solar radiation on that day acquired by the acquisition unit 110, and the learning data stored in the memory unit (past power generation prediction values ​​and actual power generation values, etc.), the calculation unit 120 predicts the amount of power generated every 30 minutes on that day. In this way, the calculation unit 120 acquires the power generation prediction values ​​every 30 minutes on that day (see Figure 3(a)). The acquired power generation prediction values ​​are stored in the memory unit of the calculation unit 120.

[0036] After performing the processing in step S11, the arithmetic unit 120 proceeds to step S12.

[0037] In step S12, the calculation unit 120 determines whether or not output control may be implemented on that day. If there is a possibility of output control being implemented, the power company will make an announcement to that effect, and the calculation unit 120 makes this determination based on this information from the power company.

[0038] If the calculation unit 120 determines that output control may be implemented the following day (YES in step S12), it proceeds to step S13. On the other hand, if the calculation unit 120 determines that output control may not be implemented the following day (NO in step S12), it terminates the predictive control shown in Figure 2.

[0039] In step S13, the calculation unit 120 performs a correction process to correct the predicted power generation value for the day obtained in step S11. In this process, the calculation unit 120 performs the correction process based on past data (predicted power generation value and actual power generation value) stored in the memory unit of the calculation unit 120.

[0040] Specifically, the calculation unit 120 obtains the ratio of the actual power generation value on a given day to the predicted power generation value on a given day of output control as the output rate. Here, the "predicted power generation value" is not the corrected predicted power generation value described later, but the uncorrected predicted power generation value initially obtained in the processing in step S11. The calculation unit 120 obtains the output rate every 30 minutes for the given day of operation. The data for the given day of operation (predicted power generation value and actual power generation value) is selected based on calendar information, and it is preferable, for example, to select data from past output control operation dates for the same day of the week as the prediction target day (the current day). Furthermore, it is even more preferable to select data from the most recent day among the data for the same day of the week.

[0041] The calculation unit 120 corrects the power generation prediction value obtained in step S11 using the acquired output rate. Specifically, the calculation unit 120 predicts the time period in which output control will be implemented based on the output rate of each time period (frame). Specifically, the calculation unit 120 sets the time period in which the output rate is below a predetermined threshold as the time period in which output control is predicted to be implemented (hereinafter referred to as the "predicted target time period"). In this embodiment, the predetermined threshold is assumed to be 100%.

[0042] The calculation unit 120 corrects the predicted power generation value for each time slot within the time frame to be predicted by multiplying the predicted power generation value by the output rate (percentage [%] value × 1 / 100). As shown in Figure 3(b), for example, if the output rate is less than 100% between 8:00 and 15:30, the calculation unit 120 corrects the predicted power generation value for each time slot from 8:00 to 15:30. In this way, the corrected predicted power generation value can be obtained by the deadline for bidding on the wholesale electricity market (10:00 the previous day) (see Figure 4(a)).

[0043] In this way, the calculation unit 120 performs correction processing for the time period to be predicted. On the other hand, the calculation unit 120 does not perform correction processing for time periods other than the time period to be predicted. This simplifies the correction processing.

[0044] After performing the processing in step S13, the arithmetic unit 120 proceeds to step S14.

[0045] In step S14, the calculation unit 120 determines whether the electricity market price (JPEX price) has been determined. The electricity market price is made public at 10:00 on the bidding day. The calculation unit 120 makes this determination based on the information about the electricity market price obtained via the acquisition unit 110.

[0046] If the calculation unit 120 determines that the electricity market price has been determined (YES in step S14), it proceeds to step S15. On the other hand, if the calculation unit 120 determines that the electricity market price has not been determined (NO in step S14), it repeats the process in step S14. In other words, the calculation unit 120 does not proceed with processing until it determines YES in step S14.

[0047] In step S15, the calculation unit 120 resets the forecast target time period based on the finalized value of the electricity market price. Specifically, the calculation unit 120 assumes that the time period in which the finalized value of the electricity market price is less than or equal to a predetermined price (for example, the lowest market price of 0.01 yen / kWh) is a time period in which output control is likely to be implemented, and resets that time period to the forecast target time period.

[0048] For example, if the electricity market price between 8:30 and 16:00 is 0.01 yen / kWh or less, the calculation unit 120 resets the time period from 8:30 to 16:00 as the forecast period. By resetting the forecast period in this way, the forecast period is expanded to include the time period in which the output rate is 100% or more (the 16:00 slot). In this case, the calculation unit 120 uses the output rate of the 15:30 slot, which is the closest to the 16:00 slot among the time periods from 8:00 to 15:30 in which the output rate is less than 100%, to perform a correction process for the 16:00 slot. Specifically, the calculation unit 120 corrects the predicted power generation value for the 16:00 slot by multiplying the predicted power generation value for the 16:00 slot by the output rate of the 15:30 slot (see Figure 4(b)).

[0049] Furthermore, for time slots where the output rate is less than 100% and the electricity market price is not 0.01 yen / kWh or less (the 8 o'clock time slot), these may be included in the forecast period or excluded from the forecast period. In this embodiment, the 8 o'clock time slot where the electricity market price is not 0.01 yen / kWh or less is included in the forecast period.

[0050] After performing the processing in step S15, the arithmetic unit 120 proceeds to step S16.

[0051] In step S16, the calculation unit 120 determines whether or not an output control instruction has been issued. If output control is to be performed, the power company will announce the output control instruction after 4 PM (4 PM to 6 PM) on the day before the day to be controlled. If the calculation unit 120 obtains information via the acquisition unit 110 that an output control instruction has been announced, it determines that an output control instruction has been issued.

[0052] If the calculation unit 120 determines that there is an output control instruction (YES in step S16), it proceeds to step S18. On the other hand, if the calculation unit 120 determines that there is no output control instruction yet (NO in step S16), it proceeds to step S17.

[0053] In step S17, the calculation unit 120 determines whether the scheduled time for the output control instruction (for example, 6pm the previous day) has passed.

[0054] If the calculation unit 120 determines that the scheduled time for the output control instruction has passed (YES in step S17), it decides that the output control instruction will not be implemented and terminates the predictive control shown in Figure 2. In this case, the predicted power generation value stored in the memory unit of the calculation unit 120 is not the predicted power generation value corrected in steps S13 and S15, but the uncorrected predicted power generation value obtained in step S11 (predicted power generation value based on solar radiation, etc.). On the other hand, if the calculation unit 120 determines that the scheduled time for the output control instruction has not passed (NO in step S17), it returns to step S16. In other words, the calculation unit 120 does not proceed with processing until YES is obtained in step S17.

[0055] In step S18, the calculation unit 120 readjusts the predicted power generation value based on the output control instruction. Specifically, the output control instruction publishes information on the time period subject to output control and the amount of output control (output control amount) for the entire power supply area. The amount of output control due to the output control instruction is shown within a range that has a minimum and maximum value. The calculation unit 120 corrects the predicted power generation value only for the time period subject to output control due to the output control instruction.

[0056] In other words, if the time period subject to output control is reduced compared to the prediction target time period reset in step S15, the calculation unit 120 restores the predicted power generation value to the uncorrected predicted power generation value obtained in step S11 for the time slots that are no longer subject to output control. On the other hand, if the time period subject to output control is expanded compared to the prediction target time period reset in step S15, the calculation unit 120 corrects the predicted power generation value for the time slots that have newly become subject to output control in the same manner as in steps S13 and S15.

[0057] Furthermore, the calculation unit 120 corrects the predicted power generation value for the prediction period based on the amount of output control (output control amount) in the entire power supply area due to the output control instruction. Specifically, the calculation unit 120 corrects the output rate by the ratio of the maximum value of the output control amount on past output control implementation days (the target days for calculating the output rate in step S13) to the maximum value of the output control amount on the prediction target day. For example, if the maximum value of the output control amount on past output control implementation days is 475 [kWh] and the maximum value of the output control amount on the prediction target day is 467 [kWh], the corrected output rate is calculated as "output rate before correction × (467 / 475)". Then, the calculation unit 120 corrects the predicted power generation value again by multiplying the predicted power generation value corrected in step S13 or step S15 by the corrected output rate (see Figure 5).

[0058] After performing the processing in step S18, the arithmetic unit 120 proceeds to step S19.

[0059] In step S19, the calculation unit 120 corrects the training data. Here, the storage unit of the calculation unit 120 stores training data for predicting the amount of power generated (step S13). The training data includes past power generation prediction values ​​and actual power generation values ​​(past data of power generation prediction values ​​and actual power generation values).

[0060] In this case, during the time periods in which output control is implemented, the actual power generation value is lower than the amount of solar radiation. Therefore, if these actual power generation values ​​are included in the training data, the correlation between solar radiation and power generation will be impaired. Specifically, if a predictive model is constructed using data on actual power generation during the time periods in which output control is implemented, the predicted power generation value for time periods without output control may be underestimated compared to the actual power generation value.

[0061] Therefore, for the time periods in which output control is implemented, the calculation unit 120 stores the uncorrected predicted power generation values ​​(see step S11), calculated from solar radiation and other factors, as training data in the memory unit, instead of the actual power generation values. This improves the accuracy of the prediction model (machine learning model) for predicting power generation, and consequently, further improves the accuracy of power generation prediction using the prediction model. Specifically, it prevents the predicted power generation values ​​for time periods in which output control is not implemented from being calculated as smaller than the actual power generation values.

[0062] After performing the processing in step S19, the calculation unit 120 terminates the predictive control shown in Figure 2.

[0063] As described above, in the power generation forecasting system 100 according to this embodiment, if there is a possibility that output control will be implemented, the predicted power generation value based on solar radiation, etc., is corrected using data (output rate) from past days when output control was implemented. This reduces the discrepancy between the predicted power generation value and the actual power generation value during the time period subject to output control, and consequently improves the accuracy of predicting the amount of power generated when output control is implemented.

[0064] Furthermore, after the wholesale electricity market auction, the system predicts the time periods when output control is most likely to be implemented based on the finalized electricity market price, and resets the prediction target time period. This improves the accuracy of the prediction of power generation during output control.

[0065] Furthermore, when output control instructions are announced, the amount of power generated is readjusted based on the target time period and the amount of output control specified in the instructions, thus improving the accuracy of power generation predictions when output control is implemented.

[0066] Furthermore, in the power generation prediction system 100 according to this embodiment, for the time period subject to output control, the system stores predicted power generation values ​​calculated from solar radiation, etc., rather than actual power generation values ​​under output control, as learning data for predicting power generation. This allows for further improvement in the accuracy of power generation prediction using the prediction model.

[0067] As described above, the power generation prediction system 100 according to this embodiment is A power generation prediction system 100 that predicts the amount of power generated by a power generation unit 200 that can generate power using natural energy, A power generation prediction value acquisition unit (calculation unit 120) predicts the amount of power generated based on predetermined information and acquires the predicted amount of power generated as a power generation prediction value (step S11), A storage unit (calculation unit 120) stores the predicted power generation value and the actual power generation value which is the actual amount of power generated (step S19), The output rate acquisition unit (calculation unit 120) acquires the ratio of the actual power generation value on a given day to the predicted power generation value on a given day of past output control implementation as the output rate (step S13), If there is a possibility that output control will be implemented on the day of the forecast (YES in step S12), the correction unit (calculation unit 120) performs a correction process to correct the predicted power generation value using the output rate (step S13), It is equipped with the following features.

[0068] This configuration can improve the accuracy of power generation forecasts. Specifically, this reduces the discrepancy between the predicted power generation value and the actual power generation value on the day output control is implemented.

[0069] Furthermore, the correction unit, The correction process is performed based on calendar information.

[0070] This configuration allows for improved accuracy in predicting power generation. Specifically, for example, correction processing can be performed using predicted and actual power generation values ​​for days close to the target date in terms of the trend in power generation (for example, the most recent day of the week immediately preceding the target date), thereby further improving the accuracy of power generation forecasts.

[0071] Furthermore, the correction unit, Based on the output rate for each predetermined time period, the target time period in which the output control will be implemented is predicted, and the predicted target time period is set as the predicted target time period. The correction process is performed for the time period to be predicted, but not for time periods other than the time period to be predicted.

[0072] This configuration simplifies the correction process.

[0073] Furthermore, the correction unit, After the electricity market price is determined, the forecast target time period is reset based on the time period in which the determined electricity market price is below a predetermined price (step S15).

[0074] This configuration allows for accurate prediction of the time period during which output control is required. Specifically, since it can be assumed that output control is likely to be implemented during periods when the electricity market price is below a predetermined price (for example, the lowest market price of 0.01 yen / kWh), the time periods subject to output control can be predicted with high accuracy.

[0075] Furthermore, the correction unit, If the prediction target time period is expanded as a result of the reset, and one time period in which the output rate is equal to or greater than a predetermined threshold (e.g., 100%) is included in the prediction target time period, the correction process for that time period is performed using the output rate of the time period closest to that time period among the time periods in which the output rate is less than the predetermined threshold (step S15).

[0076] This configuration improves the accuracy of power generation forecasts within an expanded time period (a single time period). Specifically, by using the power output rate of a time period that is likely to have a similar trend in power output rate to the expanded forecast period (the first time period), the accuracy of power generation forecasts for the expanded forecast period (the first time period) can be improved.

[0077] Furthermore, the correction unit, If an output control instruction is issued that includes information on the target time period and the amount of output control (YES in step S16), the correction process is re-executed (step S18) based on the target time period and the amount of output control specified in the output control instruction.

[0078] This configuration allows for improved accuracy in predicting power generation. Specifically, by re-performing the correction process based on highly accurate output control instruction information, the accuracy of power generation predictions can be further improved.

[0079] Furthermore, the aforementioned predetermined information includes: The training data includes past power generation performance values, The aforementioned storage unit is For the time period during which the output control described above was implemented, the uncorrected power generation forecast value initially acquired by the power generation forecast value acquisition unit is stored as the learning data instead of the actual power generation value.

[0080] This configuration allows for a further improvement in the accuracy of power generation predictions. Specifically, this prevents the predicted power generation value during periods when output control is not implemented from being lower than the actual power generation value.

[0081] Although embodiments of the present invention have been described above, the present invention is not limited to the above configuration, and various modifications are possible within the scope of the invention as described in the claims.

[0082] For example, in this embodiment, the power generation unit 200 is designed to generate electricity using sunlight, but any unit capable of generating electricity using natural energy sources would suffice.

[0083] Furthermore, in this embodiment, the past output control implementation dates (step S13) for calculating the output rate are selected based on calendar information. For example, data from past output control implementation dates that are the same day of the week as the prediction target date (the current day) are selected, but this is not limited to this. For example, if the prediction target date is a weekday, a weekday may be selected from past output control implementation dates, and if the prediction target date is a Saturday, Sunday, or public holiday, a Saturday, Sunday, or public holiday may be selected from past output control implementation dates. [Explanation of symbols]

[0084] 100 Power Generation Forecasting Systems 120 Arithmetic section 200 Power Generation Unit

Claims

1. A power generation prediction system that predicts the amount of power generated by a power generation unit that can generate power using natural energy, A power generation prediction value acquisition unit that predicts the amount of power generated based on predetermined information and acquires the predicted amount of power generated as a power generation prediction value, A storage unit that stores the predicted power generation value and the actual power generation value which is the actual amount of power generated, An output rate acquisition unit that acquires the ratio of the actual power generation value on a given day to the predicted power generation value on a given day of output control implementation as the output rate, If there is a possibility that output control will be implemented on the day of the forecast, the correction unit performs a correction process to correct the predicted power generation value using the output rate, A power generation forecasting system equipped with the following features.

2. The correction unit, The correction process is performed based on the calendar information. The power generation prediction system according to claim 1.

3. The correction unit, Based on the output rate for each predetermined time period, the target time period in which the output control will be implemented is predicted, and the predicted target time period is set as the predicted target time period. The correction process is performed for the time period to be predicted, but not for time periods other than the time period to be predicted. The power generation prediction system according to claim 1.

4. The correction unit, After the electricity market price is determined, the forecast target time period is reset based on the time period during which the determined electricity market price is below a predetermined price. The power generation prediction system according to claim 3.

5. The correction unit, If the prediction target time period expands as a result of the reset, and one time period in which the output rate is equal to or greater than a predetermined threshold is included in the prediction target time period, the correction process for that time period is performed using the output rate of the time period closest to that time period among the time periods in which the output rate is less than the predetermined threshold. The power generation prediction system according to claim 4.

6. The correction unit, When an output control instruction is issued that includes information on the target time period and the amount of output control, the correction process is re-executed based on the target time period and the amount of output control specified in the output control instruction. The power generation prediction system according to claim 1.

7. The aforementioned predetermined information includes: The training data includes past power generation performance values, The aforementioned storage unit is For the time period during which the output control described above was implemented, the uncorrected power generation forecast value initially acquired by the power generation forecast value acquisition unit is stored as the learning data instead of the actual power generation value. A power generation prediction system according to any one of claims 1 to 6.

Images