Bidding curve prediction apparatus, bidding curve prediction method, and program

Abstract

To accurately predict future bidding curves.SOLUTION: A bidding curve prediction apparatus comprises: an acquisition unit for acquiring operating state information of each power generation facility that is bidding in the electricity market; an estimation unit for estimating bidding information, which is a bidding price and a bidding volume for each power generation facility, based on a relationship between a bidding curve that indicates the relationship between the bidding price and the bidding volume and the operating state information; a correction unit that corrects the bidding curve based on the bidding information estimated by the estimation unit; and a prediction unit for predicting the bidding curve on a prediction target date using the bidding curve corrected by the correction unit.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to a bidding curve prediction device, a bidding curve prediction method, and a program. [Background technology]

[0002] With the advancement of electricity deregulation, electric utilities are now able to procure electricity through electricity trading at the wholesale electricity exchange. The Japan Electric Power Exchange (JEPX), a market for such electricity trading, includes a day-ahead market (spot market) for trading electricity for next-day delivery. In the spot market, bids are submitted for bids and buys for 48 products, divided into 48 slots, each representing a 30-minute period. The price and quantity of buy and sell bids are bid for each slot. The spot market employs a blind single-price auction as the contract method. In a blind single-price auction, the contract price is determined by computer calculation of the equilibrium point where the buy and sell bid curves intersect. Currently, the JEPX website publishes the actual bidding curves, allowing users to understand the trading history of past electricity transactions. However, bidding trends before the contract price is determined are not disclosed, preventing bidders from understanding the actual bidding curve on the target day. Therefore, predicting future bidding curves is necessary when considering procurement plans and bidding strategies.

[0003] A bidding curve is formed by stacking bid quantities, which represent the amount of electricity bid, in ascending or descending order of bid price. For example, if a specific power generator is shut down due to inspection or an accident, the shape of the selling bid curve changes due to the absence of selling bids corresponding to that generator. Alternatively, in the same situation, in order to procure the amount of electricity previously consumed by that generator from the market, buying bids corresponding to the amount of electricity consumed by that generator are generated, causing the shape of the buying bid curve to change. Therefore, when predicting a future bidding curve, it is not sufficient to simply refer to the performance of past bidding curves; it is necessary to correct the bidding curve to reflect the operating status of the generator. Patent Document 1 discloses a technology for correcting a bidding curve by taking into account fluctuations in the amount of electricity generated by variable power sources such as solar power generation. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2022-041024 Summary of the Invention [Problem to be solved by the invention]

[0005] However, in the conventional technology, it is assumed that the entire amount of electricity generated by variable power sources is bid at the minimum bid price, for example, 0.01 yen / kWh, when bidding on the market. This is because, due to the nature of variable power sources, it is considered necessary to prioritize contracts over electricity generated by conventional power generation facilities using hydroelectric, thermal, and nuclear power. In other words, while electricity from variable power sources is considered to be bid at a uniform price, the conventional technology cannot be applied to correct the bidding curve according to the operating status of conventional power generation facilities, whose bidding prices vary depending on the power generation facility. This has led to the problem of difficulty in accurately predicting future bidding curves.

[0006] The present disclosure has been made in consideration of the above circumstances, and provides a bid curve prediction device, a bid curve prediction method, and a program that are capable of accurately predicting future bid curves. [Means for solving the problem]

[0007] The present invention has been made to solve the above-mentioned problems, and one aspect of the present disclosure is a bid curve prediction device that includes an acquisition unit that acquires operating state information of each power generation facility that will bid in an electricity market, an estimation unit that estimates bid information that is the bid price and bid volume for each power generation facility based on the relationship between a bid curve that represents the relationship between the bid price and the bid volume and the operating state information, a correction unit that corrects the bid curve based on the bid information estimated by the estimation unit, and a prediction unit that predicts the bid curve on a target prediction date using the bid curve corrected by the correction unit.

[0008] Another aspect of the present disclosure is a bid curve prediction method performed by a bid curve prediction device that is a computer, in which an acquisition unit acquires operating state information of each power generation facility that will bid in an electricity market, an estimation unit estimates bid information that is a bid price and bid volume for each power generation facility based on the relationship between a bid curve that represents the relationship between a bid price and a bid volume and the operating state information, a correction unit corrects the bid curve based on the bid information estimated by the estimation unit, and a prediction unit predicts the bid curve on a prediction date using the bid curve corrected by the correction unit.

[0009] Another aspect of the present disclosure is a program that causes a bid curve prediction device, which is a computer, to acquire operating state information of each power generation facility that will bid in the electricity market, estimate bid information that is the bid price and bid volume for each power generation facility based on the relationship between a bid curve that represents the relationship between the bid price and the bid volume and the operating state information, correct the bid curve based on the estimated bid information, and predict the bid curve on a target prediction date using the corrected bid curve. [Effects of the Invention]

[0010] According to this disclosure, future bidding curves can be predicted with high accuracy. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a diagram illustrating an example of the configuration of a bid curve prediction system 10 according to an embodiment. [Figure 2] FIG. 2 is a block diagram showing an example of the configuration of a prediction device 30 according to an embodiment. [Figure 3] FIG. 3 is a diagram illustrating an example of the configuration of information indicating the operating state of a power generation facility according to an embodiment. [Figure 4] FIG. 10 is a diagram illustrating an example of a bidding curve according to an embodiment. [Figure 5] FIG. 10 is a diagram illustrating an example of bid information according to an embodiment. [Figure 6]FIG. 10 is a diagram illustrating an example of correction of a bid curve performed by the prediction device 30 in the embodiment. [Figure 7] FIG. 10 is a diagram illustrating an example of correction of a bid curve performed by the prediction device 30 in the embodiment. [Figure 8] FIG. 10 is a diagram illustrating an example of correction of a bid curve performed by the prediction device 30 in the embodiment. [Figure 9] FIG. 10 is a diagram illustrating an example of correction of a bid curve performed by the prediction device 30 in the embodiment. [Figure 10] 10 is a flowchart showing the flow of processing performed by a prediction device 30 in an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0013] 1 is a diagram showing an example of the configuration of a bid support system 1 according to an embodiment. The bid support system 1 includes, for example, a bid curve prediction system 10 and a viewing terminal 50.

[0014] The bidding curve prediction system 10 is a system for predicting a bidding curve, for example, a system for predicting a bidding curve in an electricity trading market using a blind single-price auction system. In the following, the present embodiment describes an example in which the bidding curve prediction system 10 predicts a bidding curve used to determine a contract price in the JEPX spot market. However, the present invention is not limited to this example. The bidding curve prediction system 10 can be applied to any system that predicts a bidding curve in a mechanism for determining a transaction price or the like based on a bidding curve. The viewing terminal 50 is a terminal device that displays the predicted value of the bidding curve predicted by the bidding curve prediction system 10. A user views the predicted value displayed on the viewing terminal 50 and, for example, makes a bid taking the predicted value into consideration. The bidding curve prediction system 10 includes, for example, an external server 20, a prediction device 30, and an internal server 40.

[0015] The external server 20 is an external server that provides a web service. The web service here is a service that provides information related to power supply and demand, such as information related to bidding and information indicating the operating status of power generation facilities. For example, bidding information can be obtained from trading market data published by JEPX. Bidding information includes information on the contract volume and contract price for each area of ​​48 product blocks for each past electricity delivery date in the spot market, as well as the contract volume and contract price for the entire market, and bidding curves for each area of ​​48 product blocks. The bidding curves include both selling bid curves and buying bid curves. The operating status of a power generation facility is information that indicates the amount of output authorized for the facility, whether the facility is operating, and if so, whether it is subject to operational restrictions. For example, JPEX operates a power generation information disclosure system (HJKS) separate from its electricity trading system. The HJKS requires power generation companies to register information regarding unplanned or planned shutdowns of power generation units with an authorized output of 100,000 kilowatts or more, and the registered information is made public. The external server 20 is a service that provides trading market data published by, for example, JEPX, and also a service that provides information indicating the operating status of power generation facilities such as HJKS.

[0016] The internal server 40 is a database that stores information about bidding obtained through the web service and information indicating the operating status of the power generation facility. For example, trading market data published by JEPX can be used as information about bidding. For example, information indicating the operating status of the power generation facility published by HJKS can be used as information indicating the operating status of the power generation facility. The internal server 40 obtains information about bidding and information indicating the operating status of the power generation facility from the external server 20, and stores the obtained information.

[0017] The prediction device 30 predicts the contract price on the target prediction date. The prediction device 30 acquires information indicating the operating state of the power generation facility from the internal server 40. The prediction device 30 also acquires the performance of information related to past bids from the internal server 40. The prediction device 30 predicts the contract price on the target prediction date based on the operating state of the power generation facility acquired from the internal server 40 and the performance of information related to past bids. The prediction device 30 outputs the predicted contract price on the target prediction date to the viewing terminal 50.

[0018] In order to predict the contract price on the target prediction date, the prediction device 30 first estimates the bid volume and bid price to be bid on the market as a sell bid for each power generation facility, and then corrects the bidding curve based on the estimated bid volume and bid price.

[0019] As described above, when a specific generator is stopped due to inspection work, an accident, etc., the shape of the sell bid curve changes due to the absence of a sell bid corresponding to that generator. Also, in a situation where the amount of power generated by a power generation facility does not change significantly, it is thought that the bid quantity and bid price submitted to the market as a sell bid corresponding to that power generation facility will not change significantly.

[0020] Taking advantage of this property, the prediction device 30 of this embodiment estimates the bid quantity and bid price to be bid on the market as a sell bid for each power generation facility based on the actual performance of the bid curve and the operating status of the power generation facility on the target date of the bid curve.

[0021] For example, suppose a first power generation facility that was operating on a first target day stops on a second target day. In this case, for example, all or part of the electricity generated by the first power generation facility on the first target day is sold to the spot market of the previous day, and on the second target day, no electricity is generated by the first power generation facility, so no sell bids are made to the spot market. In such a situation, if the first power generation facility is the only power generation facility whose operating state changed between the first and second target days, it is possible to estimate that the difference between the sell bid curves for the first and second target days represents the bid quantity and bid price of the sell bid corresponding to the first power generation facility.

[0022] The prediction device 30 then corrects the bidding curve based on the estimated bid volume and bid price for each power generation facility. For example, by correcting the actual bidding curve, the prediction device 30 can generate a bidding curve assuming that the operating state of the power generation facility was in a specific state on the past bidding date corresponding to the bidding curve. Alternatively, the prediction device 30 can first predict a bidding curve for a target prediction date assuming that the operating state of the power generation facility is in a specific state, and then generate a bidding curve corresponding to the actual operating state of the power generation facility scheduled for the predicted target date. This makes it possible to identify discontinuous changes in the electricity trading price and to estimate an electricity trading price corresponding to an operating state of a generator that did not exist in the past. Therefore, it is expected to be effective in accurately predicting future bidding curves.

[0023] The following describes an example in which the "selling bid curve" is corrected, but the present invention is not limited to this. The same concept can be applied to a case in which the "buying bid curve" is also the subject of correction. The same concept can also be applied to a case in which both the "selling bid curve" and the "buying bid curve" are the subject of correction.

[0024] A specific method by which the prediction device 30 corrects the bid curve will be described with reference to Fig. 2. Fig. 2 is a block diagram showing an example configuration of the prediction device 30 in an embodiment. The prediction device 30 includes, for example, an operating state acquisition unit 301, a bid curve acquisition unit 302, a bid information estimation unit 303, a bid curve correction unit 304, a reference bid curve generation unit 305, and a bid curve prediction unit 306.

[0025] The operating state acquisition unit 301 acquires the operating state of the power generation equipment. The operating state acquisition unit 301 acquires information indicating the operating state of the power generation equipment stored in the internal server 40, and stores the operating state of the power generation equipment, such as whether or not the power generation equipment is operating, operating constraints, etc., in a storage unit (not shown) in association with the date and time.

[0026] FIG. 3 is a diagram illustrating an example of the configuration of information indicating the operating state of a power generation facility in an embodiment. As illustrated in FIG. 3, the information indicating the operating state of the power generation facility includes information corresponding to items such as a power generation facility ID, a date, authorized output, a category, a type, and a reduction amount. The power generation facility ID is identification information that can uniquely identify the power generation facility, such as a power plant code. The date is a date corresponding to the operating state of the power generation facility. The authorized output is the amount of power generation authorized for the power generation facility. The category, type, and reduction amount each indicate the operating state. The category is a category of the operating state, such as operation, planned shutdown, unplanned shutdown, and output reduction. The type is a type corresponding to the category. For example, when the operating state is classified as a planned shutdown, the type indicates the reason for the shutdown, such as shutdown due to constraints on transmission lines or equipment failure, or the like. When the operating state is classified as output reduction, the type indicates the reason for the output reduction, such as output reduction due to constraints on transmission lines or the like, or output reduction due to other reasons, or the like. The reduction amount is the amount of reduction when the operating state is classified as output reduction.

[0027] In this diagram, the power generation facility identified by power generation facility ID (D0001) has a licensed output of 360 MW on the date (YYYY / MM / DD) and is operating normally. The power generation facility identified by power generation facility ID (D0002) has a licensed output of 600 MW on the date (YYYY / MM / DD) and is scheduled to be shut down due to restrictions such as transmission lines. The power generation facility identified by power generation facility ID (D0003) has a licensed output of 500 MW on the date (YYYY / MM / DD) and is operating at a reduced output of 115 MW due to restrictions such as transmission lines.

[0028] Note that the information indicating the power status of the power generation facility is not limited to the example shown in this figure, and may include information indicating the area where the power generation facility is installed, the power generation company, the name of the power generation facility, the power generation type, etc. The power generation type is information indicating the type of power generation, such as hydroelectric power generation, thermal power generation, nuclear power generation, and power generation using new energy sources such as solar power generation and wind power generation. Furthermore, if the operating status of the power generation facility is "stopped," the date of the stoppage and whether or not there is an outlook for recovery may be displayed as information indicating the power status of the power generation facility. In addition, in the example of this figure, the operating status of the power generation facility is shown for each date, but is not limited to this. For example, a period during which the power generation facility is operating in a non-normal manner, such as being stopped or operating at a reduced output, may be shown as information indicating the power status of the power generation facility.

[0029] 2, the bid curve acquisition unit 302 acquires the performance of the bid curve. The bid curve acquisition unit 302 acquires information indicating the performance of the sell bid curve among the bid curves stored in the internal server 40, and stores the information in a storage unit (not shown) in association with information indicating the date the bid was made, the block that was the subject of the bid, the state of market fragmentation on the date the bid was made, etc.

[0030] Here, in a bid for electricity trading, a bid price, a bid quantity, and a bidding area are specified. A bidding area is an area where electricity is generated when selling electricity, and an area where electricity is consumed when purchasing electricity. Examples of areas include nine regions: Hokkaido, Tohoku, Tokyo, Chubu, Hokuriku, Kansai, Chugoku, Shikoku, and Kyushu. Electricity is transmitted between areas via transmission lines that connect the areas, called interconnection lines. Interconnection lines have an upper limit on the amount of electricity they can transmit, and they cannot transmit more than a certain amount. As a result, when one area purchases electricity from another area, it cannot purchase more than a certain amount.

[0031] Market segmentation occurs when a market that should be one is split into two due to a lack of transmission capacity on the interconnection line between two areas. When market segmentation occurs, the bidding curves differ for each area, resulting in different contract prices for each area. In this case, for example, the contract price in an area where supply and demand is tighter will be higher than the contract price in the other area.

[0032] The bidding curve acquisition unit 302 acquires a bidding curve to be corrected by the bidding curve correction unit 304, which will be described later. In other words, when the "selling bid curve" is corrected by the bidding curve correction unit 304, the bidding curve acquisition unit 302 acquires information indicating the performance of the "selling bid curve." On the other hand, when the "buying bid curve" is corrected by the bidding curve correction unit 304, the bidding curve acquisition unit 302 acquires information indicating the performance of the "buying bid curve." When both the "selling bid curve" and the "buying bid curve" are corrected by the bidding curve correction unit 304, the bidding curve acquisition unit 302 acquires information indicating the performance of both the "selling bid curve" and the "buying bid curve."

[0033] Fig. 4 is a diagram showing an example of a bidding curve in an embodiment. The horizontal axis of Fig. 4 indicates the bid volume, and the vertical axis indicates the bid price. Fig. 4 shows an example of a selling bid curve. As shown in the example of this diagram, the selling bid curve is formed by stacking the bid volumes in ascending order of bid price. In Figure 4, the selling bid curve is shown as a "straight diagonal line" with an upward slope, but this is a general representation of the bidding curve, and an actual bidding curve does not have a "straight diagonal line." As shown in the enlarged image, when observing the bidding curve locally, "horizontal lines" corresponding to the bid quantity and bid price are arranged in ascending order of price, and adjacent "horizontal lines" are generally connected by "vertical lines" corresponding to the price difference between them, forming a stepped curve. The same is true for the bidding curves shown in Figures 6 to 9, which will be described later.

[0034] 2, the bid information estimation unit 303 estimates bid information. The bid information is information indicating, for each power generation facility, an estimated value of the bid amount and bid price placed as a selling bid corresponding to the power generation facility.

[0035] It is known that power demand varies depending on the day of the week, time of day, etc. Power generation companies are expected to make power selling bids taking such changes in power demand into consideration. For this reason, the bid information estimation unit 303 estimates bid information for each day of the week, each time slot, and a combination of these. Note that the day of the week here may include not only the day of the week, but also whether it is a weekday or a holiday, whether a day is a day on which a specific event is held, etc.

[0036] The bid information estimation unit 303 estimates bid information based on the performance of the bid curve and the operating status of the power generation facility on the target date of the bid curve. For example, the bid information estimation unit 303 classifies the bid curve into categories such as day of the week, time slot, and a combination of these, and extracts differences in the operating status of each power generation facility in the target area and differences in the bid curve for each category. The bid information estimation unit 303 focuses on the power generation facility to be estimated, and extracts bid curves in which the operating states of the power generation facility are different but the conditions other than the operating states are equivalent. For example, the bid information estimation unit 303 acquires a first bid curve and a second bid curve. The first bid curve is a bid curve acquired when the operating state of the power generation facility to be estimated is in a first state (for example, an operating state). The second bid curve is a bid curve acquired when the operating state of the power generation facility to be estimated is in a second state (for example, a stopped state) and the operating conditions of the power generation facilities other than the one to be estimated are the same as those on the day the first bid curve was acquired. The bidding information estimation unit 303 compares the first bidding curve with the second bidding curve and extracts a portion where there is a large difference on the horizontal axis (bidding quantity direction). The bidding information estimation unit 303 can estimate that the magnitude of the difference on the horizontal axis in the extracted portion represents the "bid quantity" that was bid in accordance with the power generated by the power generation facility to be estimated on the first bidding curve when the power generation facility to be estimated is stopped, but was not bid on the second bidding curve. Furthermore, the bidding information estimation unit 303 can estimate that the value on the vertical axis (bidding price direction) corresponding to the "bid quantity" is the "bid price" that was bid in accordance with the power generated by the power generation facility to be estimated on the first bidding curve. The bid information estimation unit 303 estimates the bid quantity and bid price for each power generation facility by performing such estimation for each power generation facility to be estimated. The bid information estimation unit 303 stores the estimated bid information in a storage unit (not shown).

[0037] Alternatively, the bid information estimation unit 303 may estimate the bid information using a machine learning technique. For example, the bid information estimation unit 303 generates learning data in which the operating states of each power generation facility in the target area on a past bid date are associated with the actual bid curve of the target area on that bid date. The bid information estimation unit 303 generates a trained model that has learned the correspondence between the operating states of the power generation facility and the bid curve by having the learning model learn the learning data. The trained model is trained to accurately predict the bid curve as the objective variable in accordance with the operating states of the power generation facility as the explanatory variable input to the trained model by learning the correspondence between the operating states of the power generation facility and the bid curve. In the above, an example has been described in which the operating status of the power generation facility is used as an explanatory variable, but the present invention is not limited to this. The explanatory variables may also include the power demand in the target area, the nationwide power demand, and the power output of solar power generation or the like that meets this demand. Furthermore, factors that indirectly determine these, such as temperature and solar radiation, may also be used as explanatory variables. This allows for machine learning that takes into account conditions such as the daily power demand situation and the illumination output of solar power generation, thereby enabling more accurate prediction of the "bid price" and "bid quantity" as bidding information. Furthermore, although the above description has been given as an example of learning the operating status of power generation facilities on past "bid dates," the present invention is not limited to this. It is also possible to learn the operating status of each power generation facility in the target area not only on past "bid dates" but also on the days immediately prior to the bid date, for example, from the day before to several days before the bid date. This enables estimation that takes into account the impact of power demand on or before the day before the bid date on the bid volume and bid price on the bid date.

[0038] The bid information estimation unit 303 causes the trained model to estimate a bidding curve (first bidding curve) when all power generation facilities expected to operate in the target category are operating. Also, the trained model estimates a bidding curve (second bidding curve) when only the power generation facility to be estimated is stopped among all power generation facilities expected to operate in the target category. The bid information estimation unit 303 estimates bidding information by regarding the difference between the first bidding curve and the second bidding curve as the bid quantity and bid price submitted as a sell bid corresponding to the power generation facility to be estimated. When estimating bid information using machine learning, the process of extracting the difference between the first and second bid curves estimated by the trained model may be automated. In this case, the bid information estimation unit 303 extracts the difference between the two bid curves using a technique such as pattern matching.

[0039] FIG. 5 is a diagram showing an example of bidding information in an embodiment. As shown in FIG. 5, the bidding information includes information corresponding to items such as a power generation facility ID, date, bid volume, and bid price. The power generation facility ID is identification information that can uniquely identify a power generation facility, such as a power plant code. The date is a date corresponding to the bidding information of the power generation facility. Instead of this date, a day of the week, a period, or a combination thereof may be set. The bid volume is an estimated value of the bid volume to be bid as a selling bid corresponding to the power generation facility. The bid price is an estimated value of the bid price to be bid as a selling bid corresponding to the power generation facility.

[0040] In this diagram, it is shown that a selling bid corresponding to the power generation facility indicated by the power generation facility ID (D0001) is estimated to be a bid of 100 MW of power at a bid price of 5 yen / kWh on date (YYYY / MM / DD). It is also shown that a selling bid corresponding to the power generation facility indicated by the power generation facility ID (D0003) is estimated to be a bid of 100 MW of power at a bid price of 5 to 7 yen / kWh on date (YYYY / MM / DD). In this way, the bid price under the bidding conditions estimated by the bid information estimation unit 303 may be estimated as a value within a certain range rather than a fixed value.

[0041] The bid curve correction unit 304 corrects the bid curve using the input information estimated by the bid information estimation unit 303. The bid curve correction unit 304 corrects the bid curve before correction so that it becomes a bid curve according to the operating state of the power generation facility. The bid curve correction unit 304 stores information in which the corrected bid curve is associated with the operating state of the power generation facility in a storage unit (not shown).

[0042] The bid curve correction unit 304 performs a first correction to correct the past bid curve so that it becomes a bid curve in which the operating state of the power generation facility is in a reference state, in order to generate learning data to be learned by the reference bid curve generation unit 305 described later.

[0043] The reference state here refers to a representative operating state of the power generation facility corresponding to the day of the week in the bidding information described above. That is, the bid curve correction unit 304 generates a bid curve corresponding to the reference state by making corrections corresponding to the reference state for each day of the week, each frame, and each combination of these. Furthermore, if the operating state of the power generation facility as the reference state varies from frame to frame, the bid curve correction unit 304 may generate a bid curve corresponding to the reference state for each frame and each frame.

[0044] In addition, the bid curve correction unit 304 performs a second correction to correct the bid curve of the reference state on the target prediction date, predicted by the bid curve prediction unit 306 described later, so that it becomes a bid curve corresponding to the actual operating state of the power generation equipment scheduled for the target prediction date.

[0045] The bid curve correction unit 304 can employ the same correction method in the first correction and the second correction. The method by which the bid curve correction unit 304 corrects the bid curve will be described with reference to Figs. 6 to 9. Figs. 6 to 9 are diagrams showing examples of correction of the bid curve performed by the prediction device 30 in the embodiment.

[0046] When making a correction to add or delete output from a specific power generation facility to or from the pre-correction bidding curve, the bid curve correction unit 304 reads and acquires bid information corresponding to the power generation facility to be corrected from a storage unit (not shown). Based on the acquired bid information, the bid curve correction unit 304 identifies a point on the pre-correction bidding curve that corresponds to the bid price of the power generation facility to be corrected as the correction start point. Here, if the bid price of the power generation facility to be corrected is not a fixed value but has a range of values, the bid curve correction unit 304 sets the correction start point to a point on the pre-correction bidding curve that corresponds to the lowest bid price of the power generation facility to be corrected.

[0047] When adding the output of a specific power generation facility to the pre-correction bidding curve, the bidding curve correction unit 304 shifts the area of ​​the bidding curve where the bid volume is larger than the correction start point along the axis (horizontal axis) indicating the bid volume in the direction of increasing the bid volume. The shift amount by which the bidding curve is shifted along the horizontal axis corresponds to the bid volume of the power generation facility to be corrected. Here, if the bid price of the power generation facility to be corrected is not a fixed value but a value within a certain range, the bid curve correction unit 304 shifts the area of ​​the bidding curve where the bid price is higher than the correction start point along the axis (vertical axis) indicating the bid price in the direction of increasing the bid price. The shift amount of the bidding curve along the vertical axis is the range of the bid prices of the power generation facility to be corrected, that is, the shift amount equivalent to the difference obtained by subtracting the minimum price from the maximum price. The bidding curve correction unit 304 sets the end of the shifted region that is closer to the correction start point as the correction end point. The bidding curve correction unit 304 performs a correction to add the output of a specific power generation facility to the bidding curve before correction by extending the bidding curve from the correction start point to the correction end point.

[0048] When deleting the output of a specific power generation facility from the pre-correction bidding curve, the bidding curve correction unit 304 shifts the bidding curve in the opposite direction to when the facility was added, by the same amount as when the facility was added. In other words, when deleting, the bidding curve correction unit 304 makes a correction opposite to when the facility was added. Specifically, the bid curve correction unit 304 sets the point on the bidding curve where the bid volume is larger than the correction start point by the bid volume of the power generation facility to be corrected as the correction completion point. The bid curve correction unit 304 shifts the area on the bidding curve where the bid volume is larger than the correction completion point along the axis (horizontal axis) indicating the bid volume in the direction of decreasing the bid volume. Furthermore, if the bid price of the power generation facility to be corrected is not a constant value but has a certain range, the bid curve correction unit 304 shifts the area on the bidding curve where the bid price is larger than the correction completion point along the axis (vertical axis) indicating the bid price in the direction of decreasing the bid price. The bid curve correction unit 304 reduces the bid curve from the correction start point to the correction completion point, thereby performing a correction to remove the output of the power generation facility from the bid curve before correction.

[0049] 6 and 7 show examples of adding a bid corresponding to the output of a specific power generation facility to a bidding curve. The examples in these figures can be applied to, for example, a correction to add the output of a power generation facility that was shut down in a past bidding curve to the bidding curve. Alternatively, the examples can be applied to a correction to add an increased output of a power generation facility that was reduced in output in a past bidding curve to the bidding curve.

[0050] FIG. 6 shows an example of a correction in which information corresponding to the power generation facility identified by the power generation facility ID (D0001) in FIG. 5, i.e., a bid volume of 100 [MW] and a bid price of 5 [yen / kWh], is added as bidding information corresponding to the power generation facility to be corrected. As shown in this figure, the bid curve correction unit 304 identifies, based on the bidding information, a point on the bidding curve at the bid price (5 [yen / kWh]) of the power generation facility to be corrected as a correction start point P1. The bid curve correction unit 304 shifts the area of ​​the bidding curve where the bid volume is larger than the correction start point P1 along the axis (horizontal axis) indicating the bid volume, by the bid volume (100 [MW]) of the power generation facility to be corrected, in the direction of increasing the bid volume. The bid curve correction unit 304 determines the end of the shifted area that is closer to the correction start point P1 as a correction end point Q1. The bid curve correction unit 304 performs the correction by extending the bidding curve from the correction start point P1 to the correction end point Q1. For example, the bid curve correction unit 304 performs correction by extending the curve portion from the correction start point P1 to the shifted correction completion point Q1 in the input curve in accordance with the bid information corresponding to the power generation facility to be corrected.

[0051] FIG. 7 shows an example of a correction that adds information corresponding to the power generation facility identified by the power generation facility ID (D0003) in FIG. 5, i.e., a bid volume of 100 [MW] and a bid price of 5 to 7 yen / kWh, as bidding information corresponding to the power generation facility to be corrected. As shown in this figure, the bid curve correction unit 304 identifies, based on the bidding information, a point on the bidding curve that is the lowest bid price (5 yen / kWh) of the power generation facility to be corrected as a correction start point P2. The bid curve correction unit 304 shifts the area of ​​the bidding curve where the bid volume is larger than the correction start point P2 along the axis indicating the bid volume (horizontal axis) in the direction of increasing the bid volume by the bid volume (100 [MW]) of the power generation facility to be corrected. Furthermore, the bid curve correction unit 304 further shifts the area of ​​the bidding curve where the bid price is larger than the correction start point P2 along the axis indicating the bid price (vertical axis) in the direction of increasing the bid price by 2 yen, which is the difference between the bid price of the power generation facility to be corrected and the difference between 5 and 7 yen. The bid curve correction unit 304 sets the end of the shifted region on the correction start point P2 side as the correction completion point Q2. The bid curve correction unit 304 performs correction by extending the bid curve from the correction start point P2 to the correction completion point Q2.

[0052] 8 and 9 show examples of deleting bids corresponding to the output of a specific power generation facility from a bidding curve. The examples in these figures can be applied to, for example, a correction for deleting the output of a power generation facility that was in operation in a past bidding curve from the bidding curve. Alternatively, the examples can be applied to a correction for deleting a reduced output of a power generation facility that was in operation in a past bidding curve from the bidding curve.

[0053] FIG. 8 illustrates an example of a correction in which information corresponding to the power generation facility identified by the power generation facility ID (D0001) in FIG. 5, i.e., a bid volume of 100 MW and a bid price of 5 yen / kWh, is deleted as the bid information corresponding to the power generation facility to be corrected. As illustrated in this figure, the bid curve correction unit 304 identifies a point on the bidding curve at the bid price (5 yen / kWh) of the power generation facility to be corrected based on the bid information as a correction start point P3. A point on the bidding curve where the bid volume is increased from the correction start point P3 by the bid volume of the power generation facility (100 MW) is determined as a correction end point Q3. The bid curve correction unit 304 shifts the area of ​​the bidding curve where the bid volume is larger than the correction end point Q3 by the bid volume of the power generation facility to be corrected (100 MW) in the direction of decreasing the bid volume. The bid curve correction unit 304 reduces the bidding curve from the correction start point P3 to the correction end point Q3. For example, the bid curve correction unit 304 reduces the curve portion from the correction start point P3 to the correction completion point Q3 in the input curve before the shift to the curve portion from the correction start point P3 to the correction completion point Q3 after the shift. This makes a correction to delete the output by the power generation facility from the bid curve before the correction. Note that this figure uses a diagram in which the relationship before and after correction in Figure 6 is reversed. As a result, the correction start point P3 and the correction completion point Q3 in the shifted area overlap. In other words, this example shows a case where the correction start point P3 and the correction completion point Q3 after the shift coincide, so there is no need to shrink the input curve. Furthermore, depending on the estimation accuracy of the bid information, it may be impossible to identify an area in the bid curve that shows a sufficient bid volume corresponding to the bid volume to be corrected, making it difficult to identify the correction start point P3 in the bid curve. In such cases, the bid curve correction unit 304 may perform exceptional processing to relax the correction conditions. For example, the bid curve correction unit 304 performs exceptional processing when it is unable to identify the correction start point P3 in the bid curve. In this case, in the exception handling, the bidding curve correction unit 304 first expands the bid price K of the power generation facility to be corrected, which is indicated in the bidding information, to the range of K±p%, where p is set according to the range allowed for the range of the bid price K. For example, the bidding curve correction unit 304 searches the pre-correction bidding curve to determine whether the amount of power indicated in the bidding information has been bid in an area indicating a price range of bid price K±p%. If there is a portion indicating the amount of bid indicated in the bidding information, the bidding curve correction unit 304 targets that portion for deletion, and sets the end with the lower price as the correction start point P3' and the end with the higher price as the correction end point Q3'. Then, the bidding curve correction unit 304 shifts the area of ​​the changed bidding curve where the bid volume is larger than the correction completion point Q3' in the direction of decreasing the bid volume by the bid volume of the power generation facility to be corrected, thereby reducing the bidding curve from the correction start point P3' to the shifted correction completion point Q3'. In this way, a correction is made to delete the output of the power generation facility from the bidding curve before correction.

[0054] FIG. 9 shows an example of a correction in which information corresponding to the power generation facility identified by the power generation facility ID (D0003) in FIG. 5, i.e., a bid volume of 100 [MW] and a bid price of 5 to 7 [yen / kWh], is deleted as the bid information corresponding to the power generation facility to be corrected. As shown in this figure, the bid curve correction unit 304 identifies, based on the bid information, a point on the bidding curve at which the lowest bid price (5 [yen / kWh]) of the power generation facility to be corrected is determined as a correction start point P4. A point on the bidding curve where the bid volume is increased from the correction start point P4 by the bid volume (100 [MW]) of the power generation facility to be corrected is determined as a correction end point Q4. The bid curve correction unit 304 shifts the area of ​​the bidding curve where the bid volume is larger than the correction end point Q4 by the bid volume (100 [MW]) of the power generation facility to be corrected in the direction of decreasing the bid volume. Furthermore, the bidding curve correction unit 304 further shifts the area of ​​the bidding curve where the bid price is higher than the correction completion point Q4 along the axis (vertical axis) indicating the bid price in the direction of decreasing the bid price by 2 yen, which is the range of the bid price of the power generation equipment to be corrected, that is, the difference between 5 and 7 yen. The bid curve correction unit 304 reduces the bid curve from the correction start point P4 to the correction completion point Q4, thereby performing a correction to remove the output by the power generation facility from the bid curve before correction. Note that this figure uses a diagram in which the relationship before and after correction in Figure 7 is reversed. As a result, the correction start point P4 and the correction completion point Q4 in the shifted area overlap. In other words, this example shows a case where the correction start point P4 and the correction completion point Q4 after the shift coincide, so there is no need to shrink the input curve. Also, the bid curve correction unit 304 may perform exception processing when the correction start point P4 cannot be identified in the bid curve before correction, similar to FIG.

[0055] Returning to the explanation of FIG. 2, the base bid curve generation unit 305 generates a bid curve in a base state. The base bid curve generation unit 305 sets a base state for each day of the week, each time slot, and each combination thereof. The base bid curve generation unit 305 also sets the operating state of the power generation facility in each base state. The base bid curve generation unit 305 acquires a bid curve corresponding to the base state by performing a first correction on a past bid curve using the bid curve correction unit 304. The base bid curve generation unit 305 generates a representative bid curve, for example, an average bid curve, from the bid curves classified and corrected according to the base state as a bid curve for that base state.

[0056] Alternatively, the base bid curve generation unit 305 may use a machine learning technique to generate a bid curve for the reference state. The base bid curve generation unit 305 obtains a bid curve corresponding to the reference state through the first correction by the bid curve correction unit 304. The base bid curve generation unit 305 generates, as learning data, data that associates the demand state for the reference state with a bid curve. The demand state here includes the power demand in the target area, the nationwide power demand, and the power generation output from solar power generation or the like that meets this demand. Furthermore, factors that indirectly determine these, such as temperature and solar radiation, may also be included.

[0057] The base bid curve generation unit 305 generates a trained model that has learned the correspondence between demand states and bid curves by having the learning model learn the learning data. The trained model is trained to be able to accurately predict a bid curve corresponding to the demand state input to the trained model by learning the correspondence between the demand state and the bid curve. The base bid curve generation unit 305 generates a bid curve obtained by inputting a demand state corresponding to a reference state into the trained model as a bid curve corresponding to the reference state. The base bid curve generation unit 305 generates bid curves corresponding to each of a plurality of reference states.

[0058] The bid curve prediction unit 306 predicts a bid curve for the target prediction date. The bid curve prediction unit 306 first generates a bid curve assuming that the state of the power generation equipment on the target prediction date is a reference state. For example, the bid curve prediction unit 306 assumes that the state of the power generation equipment on the target prediction date will be in a certain reference state, and first causes the reference bid curve generation unit 305 to generate a bid curve for that reference state. Here, the bid curve prediction unit 306 determines the reference state corresponding to the target prediction date based on the day of the week, the period, and a combination of these for the target prediction date.

[0059] Next, the bid curve prediction unit 306 acquires the operation state of the power generation facility scheduled for the prediction target date. For example, the bid curve prediction unit 306 acquires the operation state of the power generation facility scheduled for the prediction target date based on the information indicating the operation state of the power generation facility acquired by the operation state acquisition unit 301.

[0060] Then, the bid curve prediction unit 306 corrects the bid curve for the target prediction date predicted assuming the reference state so that it becomes a bid curve that corresponds to the planned operating state of the power generation facility on the target prediction date. For example, the bid curve prediction unit 306 causes the bid curve correction unit 304 to execute the second correction, thereby correcting the bid curve for the target prediction date predicted assuming the reference state so that it becomes a bid curve that corresponds to the planned operating state of the power generation facility on the target prediction date. The bid curve prediction unit 306 sets the bid curve after the second correction as the prediction result of the bid curve for the target prediction date.

[0061] The bid curve predicted by the bid curve predictor 306 in this manner can be used to predict the contract price.

[0062] FIG. 10 is a flowchart illustrating a process performed by the prediction device 30 according to an embodiment. First, the prediction device 30 acquires the results of the bid curve (step S10). Next, the prediction device 30 acquires the operating state of the power generation facility (step S11). Next, the prediction device 30 estimates bid information (step S12). Based on the bid information, the prediction device 30 corrects the past bid curve to an input curve in a reference state (step S13). As a result, the prediction device 30 generates bid curves corresponding to each reference state. Next, the prediction device 30 generates a bid curve in the reference state on the target prediction date (step S14). For example, the prediction device 30 generates a bid curve in the reference state on the target prediction date using a trained model that predicts a bid curve corresponding to the reference state based on the reference state. Then, the prediction device 30 corrects the bid curve in the reference state on the target prediction date according to the operating state of the power generation facility on the target prediction date (step S15). As a result, the prediction device 30 predicts a bid curve according to the operating state of the power generation facility on the target prediction date. In the above-described flowchart, the execution order of steps S10 and S11 may be reversed.

[0063] As described above, the prediction device 30 of this embodiment includes the operating state acquisition unit 301, the bid information estimation unit 303, the bid curve correction unit 304, and the bid curve prediction unit 306. The prediction device 30 is an example of a "bid curve prediction device." The operating state acquisition unit 301 is an example of an "acquisition unit" and acquires operating state information of each power generation facility that submits a bid in the electricity market. In the electricity market, electricity generated by each of multiple power generation facilities is traded as an object of purchase and sale. The bid information estimation unit 303 is an example of an "estimation unit" and estimates bid information, which is the bid price and bid volume for each power generation facility, based on the relationship between the operating state information and a bid curve that represents the relationship between the bid price and bid volume in the electricity market. The bid curve correction unit 304 is an example of a "correction unit" and corrects the bid curve based on the bid information estimated by the bid information estimation unit 303. The bid curve prediction unit 306 is an example of a "prediction unit" and predicts the bid curve on the prediction date using the bid curve corrected by the bid curve correction unit 304. As a result, the prediction device 30 of the embodiment can make predictions taking into account the planned operating state of the power generation facility on the prediction target date, and can predict future bid curves with high accuracy.

[0064] Furthermore, in the prediction device 30 of the embodiment, the bid curve correction unit 304 performs a first correction process. The first correction process is a process of correcting a past bid curve to a reference bid curve, which is a bid curve in a reference state in which each power generation facility is in a representative operating state, based on operating state information corresponding to the past bid curve and the bid information estimated by the bid information estimation unit 303. The bid curve prediction unit 306 predicts a bid curve in the reference state on the target prediction date using the past bid curve on which the bid curve correction unit 304 has performed the first correction process. The bid curve correction unit 304 performs a second correction process. The second correction process is a process of correcting the bid curve predicted by the bid curve prediction unit 306 based on operating state information corresponding to the target prediction date and the bid information. The bid curve prediction unit 306 predicts the bid curve on which the bid curve correction unit 304 has performed the second correction process as the bid curve on the target prediction date. As a result, the prediction device 30 of the embodiment sets a representative operating state of the power generation facility determined depending on the day of the week, etc. as a reference state, and first predicts a bidding curve assuming that the operating state of the power generation facility on the target prediction date will be the reference state using the reference state as a clue, and then corrects the predicted bidding curve according to the operating state of the power generation facility scheduled for the target prediction date. This makes it possible to predict future bidding curves with high accuracy.

[0065] Furthermore, in the prediction device 30 of the embodiment, the bid curve prediction unit 306 predicts the bid curve in the reference state on the prediction target date using a trained model. The trained model is a model trained to predict the bid curve in the reference state from the reference state by learning the correspondence between the reference state and the bid curve corrected to the reference state by the bid curve correction unit 304 performing the first correction process. As a result, the prediction device 30 of the embodiment can predict the bid curve in the reference state using the trained model, and after accurately predicting the bid curve in the reference state, can perform correction according to the operating state of the power generation facility scheduled for the prediction target date. This allows for accurate prediction of future bid curves.

[0066] In the prediction device 30 of the embodiment, the bid information is information in which a bid price indicating a fixed value is set for a bid quantity. The bid curve correction unit 304 corrects the selling bid curve as a bid curve. The bid curve correction unit 304 sets a point on the pre-correction bid curve corresponding to the bid price of the power generation facility to be corrected as a correction start point P. The bid curve correction unit 304 shifts the region of the pre-correction bid curve where the bid quantity is larger than the correction start point P by an amount equivalent to the bid quantity of the power generation facility to be corrected, in the direction of increasing the bid quantity. The bid curve correction unit 304 performs correction to add bid information for the power generation facility to be corrected by connecting the correction start point P on the pre-correction bid curve with a correction end point Q, which is the end of the shifted region on the side of the correction start point P, with a straight line. In this way, the prediction device 30 of the embodiment can directly add bid information for the power generation facility to be corrected to the pre-correction bid curve, thereby accurately correcting the curve according to the operating state of the power generation facility.

[0067] Furthermore, in the prediction device 30 of the embodiment, the bidding information is information in which a bid price having a certain range is set relative to the bid volume. The bid curve correction unit 304 corrects the selling bid curve as the bidding curve. The bid curve correction unit 304 sets a point on the bidding curve before correction that corresponds to the bid price of the power generation facility to be corrected as a correction start point P. The bid curve correction unit 304 shifts the area of ​​the bidding curve before correction where the bid volume is larger than the correction start point P by an amount equivalent to the bid volume of the power generation facility to be corrected, in the direction of increasing the bid volume, and by an amount equivalent to the difference in bid price of the power generation facility to be corrected, in the direction of increasing the bid price. The bid curve correction unit 304 performs a correction to add bid information for the power generation facility to be corrected by connecting the correction start point P on the bidding curve before correction to a correction end point Q, which is the end of the shifted area on the side of the correction start point P, with a straight line. As a result, in the prediction device 30 of the embodiment, even if a bid price with a certain range is set as the bid information, the bid information of the power generation equipment to be corrected can be directly added to the bid curve before correction, and correction can be made accurately according to the operating state of the power generation equipment.

[0068] Furthermore, in the prediction device 30 of the embodiment, the bidding information is information in which a bid price indicating a fixed value is set for a bid quantity. The bid curve correction unit 304 corrects the selling bid curve as a bidding curve. The bid curve correction unit 304 sets a point on the bidding curve before correction that corresponds to the bid price of the power generation facility to be corrected as a correction start point P. The bid curve correction unit 304 sets a point on the bidding curve before correction that is obtained by increasing the bid quantity by the bid quantity of the power generation facility to be corrected from the correction start point P. The bid curve correction unit 304 shifts the area on the bidding curve before correction where the bid quantity is larger than the correction end point Q in the direction of decreasing the bid quantity by the bid quantity of the power generation facility to be corrected. The bid curve correction unit 304 performs a correction to delete the bid information of the power generation facility to be corrected by connecting the correction start point P on the bidding curve before correction and the correction end point Q in the shifted area with a straight line. As a result, the prediction device 30 of the embodiment can directly delete the bidding information of the power generation facility to be corrected from the bidding curve before correction, and can perform accurate correction according to the operating state of the power generation facility.

[0069] In the prediction device 30 according to the embodiment, the bidding information is information in which a bid price having a certain range is set relative to the bid volume. The bid curve correction unit 304 corrects the selling bid curve as a bidding curve. The bid curve correction unit 304 sets a point on the bidding curve before correction that corresponds to the bid price of the power generation facility to be corrected as a correction start point P. The bid curve correction unit 304 sets a point on the bidding curve before correction that is obtained by increasing the bid volume by the bid volume of the power generation facility to be corrected from the correction start point P. The bid curve correction unit 304 shifts the area on the bidding curve before correction where the bid volume is larger than the correction end point Q in the direction of decreasing the bid volume by the bid volume of the power generation facility to be corrected, and also shifts the area in the direction of decreasing the bid price by an amount equivalent to the difference between the bid prices of the power generation facility to be corrected. The bid curve correction unit 304 performs a correction to delete the bid information of the power generation facility to be corrected by connecting the correction start point P on the bidding curve before correction and the correction end point Q in the shifted area with a straight line. As a result, in the prediction device 30 of the embodiment, even if a bid price with a certain range is set as the bid information, the bid information for the power generation equipment to be corrected can be directly deleted from the bid curve before correction, and correction can be made accurately according to the operating state of the power generation equipment.

[0070] In the above-described embodiment, an example was described in which a "selling bid curve" was corrected. However, when the prediction device 30 corrects a "buying bid curve," the "power generation facility" in the above-described embodiment can be replaced with "demand." Then, the operating state acquisition unit 301 acquires demand-related status information indicating the state of demand of the electricity retailer and the state of self-consumption demand. The bidding information estimation unit 303 estimates bidding information in the "selling bid curve" based on the demand-related status information.

[0071] 1 may be realized by recording a program for realizing the functions of the bid curve prediction system 10 and the prediction device 30 on a computer-readable recording medium, and reading and executing the program recorded on the recording medium into a computer system. Note that the term "computer system" here includes hardware such as an OS and peripheral devices.

[0072] Furthermore, if a WWW system is used, the "computer system" also includes the homepage provision environment (or display environment). "Computer-readable recording media" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into computer systems. Furthermore, "computer-readable recording media" also includes devices that dynamically store programs for a short period of time, such as communication lines used when transmitting programs over networks like the Internet or over communication lines like telephone lines, and devices that store programs for a fixed period of time, such as volatile memory within computer systems that serve as servers or clients. The programs may also be programs that implement some of the aforementioned functions, or may be programs that can realize the aforementioned functions in combination with programs already stored in the computer system.

[0073] Although an embodiment of the present invention has been described in detail above with reference to the drawings, the specific configuration is not limited to this embodiment, and design changes and the like are also included within the scope that does not deviate from the gist of the present invention. [Explanation of symbols]

[0074] 1. Bidding support system 10. Bid Curve Prediction System 20 External Servers 30 Prediction device (bid curve prediction device) 40 Internal Servers 50 viewing devices 301 Operating status acquisition unit (acquisition unit) 302 Bid Curve Acquisition Department 303 Bidding Information Estimation Department (Estimation Department) 304 Bid Curve Correction Unit (Correction Unit) 305 Standard Bid Curve Generation Unit 306 Bid Curve Prediction Unit (Prediction Unit)

Claims

1. an acquisition unit that acquires operating state information of each power generation facility that is bidding in the electricity market; an estimation unit that estimates bidding information, which is a bidding price and a bidding volume for each power generation facility, based on a relationship between a bidding curve that indicates a relationship between a bidding price and a bidding volume and the operating state information; a correction unit that corrects the bidding curve based on the bidding information estimated by the estimation unit; a prediction unit that predicts the bid curve on a prediction target date using the bid curve corrected by the correction unit; A bidding curve prediction device comprising:

2. the correction unit performs a first correction process to correct the past bidding curve to a reference bidding curve, which is the bidding curve in a reference state in which each power generation facility is in a representative operating state, based on the operating state information corresponding to the past bidding curve and the bidding information estimated by the estimation unit; the prediction unit predicts the bid curve in the reference state on the prediction target date using the past bid curve on which the first correction process has been performed by the correction unit; the correction unit performs a second correction process to correct the bid curve predicted by the prediction unit based on the operating state information corresponding to the prediction target date and the bid information; the prediction unit predicts the bid curve on the prediction target date, based on the bid curve on which the second correction process has been performed by the correction unit; The bidding curve prediction device according to claim 1 .

3. The prediction unit predicts the bid curve in the reference state on a prediction target date using a trained model, The learned model is a model trained to predict the bid curve in the reference state from the reference state by learning a correspondence relationship between the reference state and the bid curve corrected to the reference state by the correction unit performing the first correction process. The bidding curve prediction device according to claim 2 .

4. The bidding information is information in which a bid price indicating a fixed value is set for a bid quantity, The correction unit correcting the sell bid curve as the bid curve; A point on the bidding curve before correction that corresponds to the bid price of the power generation facility to be corrected is set as a correction start point, Shifting an area in the bidding curve before correction in which the bid volume is larger than the correction start point by an amount corresponding to the bid volume of the power generation facility to be corrected in a direction in which the bid volume increases; The bidding curve is extended from the correction start point to a correction completion point, which is the end of the shifted region on the side of the correction start point, thereby performing a correction to add the bidding information of the power generation facility to be corrected. The bidding curve prediction device according to claim 1 .

5. The bidding information is information in which a bid price having a certain range is set for a bid quantity, The correction unit correcting the sell bid curve as the bid curve; A point on the bidding curve before correction that corresponds to the lowest bid price of the power generation facility to be corrected is set as a correction start point, Shifting an area in the bidding curve before correction in which the bid volume is larger than the correction starting point by an amount equivalent to the bid volume of the power generation facility to be corrected in a direction in which the bid volume increases, and shifting an area in which the bid price increases by an amount equivalent to the difference in the bid price of the power generation facility to be corrected, The bidding curve is extended from the correction start point to a correction completion point, which is the end of the shifted region on the side of the correction start point, thereby performing a correction to add the bidding information of the power generation facility to be corrected. The bidding curve prediction device according to claim 1 .

6. The bidding information is information in which a bid price indicating a fixed value is set for a bid quantity, The correction unit correcting the sell bid curve as the bid curve; A point on the bidding curve before correction that corresponds to the bid price of the power generation facility to be corrected is set as a correction start point, a correction completion point is a point where the bid amount is increased by the bid amount of the power generation facility to be corrected from the correction start point on the bidding curve before correction; Shifting an area in the bidding curve before correction in which the bid volume is larger than the correction completion point in a direction in which the bid volume becomes smaller by the bid volume of the power generation facility to be corrected; performing a correction to delete the bidding information of the power generation facility to be corrected by reducing the bidding curve from the correction start point to the correction end point; The bidding curve prediction device according to claim 1 .

7. The bidding information is information in which a bid price having a certain range is set for a bid quantity, The correction unit correcting the sell bid curve as the bid curve; A point on the bidding curve before correction that corresponds to the lowest bid price of the power generation facility to be corrected is set as a correction start point, a correction completion point is a point where the bid amount is increased by the bid amount of the power generation facility to be corrected from the correction start point on the bidding curve before correction; Shifting an area in the bidding curve before correction where the bid volume is larger than the correction completion point in a direction to decrease the bid volume by an amount corresponding to the bid volume of the power generation facility to be corrected, and shifting an area in a direction to decrease the bid price by an amount corresponding to the difference in the bid price of the power generation facility to be corrected, performing a correction to delete the bidding information of the power generation facility to be corrected by reducing the bidding curve from the correction start point to the correction end point; The bidding curve prediction device according to claim 1 .

8. A bidding curve prediction method performed by a bidding curve prediction device that is a computer, comprising: An acquisition unit acquires operating state information of each power generation facility bidding in the electricity market, an estimation unit estimates bidding information, which is a bid price and a bid quantity for each power generation facility, based on a relationship between a bidding curve representing a relationship between a bid price and a bid quantity and the operating state information; a correction unit correcting the bidding curve based on the bidding information estimated by the estimation unit; a prediction unit predicting the bid curve on the prediction target date using the bid curve corrected by the correction unit; Bid curve prediction method.

9. The bidding curve prediction device is a computer. Obtaining operational status information of each power generation facility bidding in the electricity market; estimating bidding information, which is a bidding price and a bidding quantity for each power generation facility, based on a bidding curve representing the relationship between the bidding price and the bidding quantity and the operating state information; correcting the bidding curve based on the estimated bidding information; Using the corrected bid curve, the bid curve on the prediction target date is predicted. program.

Images