JP7647982B1 - Information processing device, information processing method, and information processing program - Google Patents

Abstract

An information processing device capable of reducing calculation costs when creating an operation plan for power equipment is provided.
[Solution] An information processing device comprising: a first acquisition unit that acquires first data including multiple past predicted values of the market price of electricity and an element related to the market price and actual values corresponding to the past predicted values; a second acquisition unit that acquires second data obtained by downsampling the first data so as to maintain a first error trend between the actual value and the predicted value of the market price and a second error trend between the actual value and the predicted value of the element; and a scenario creation unit that creates multiple scenarios for the future market price and each of the elements based on the second data and future predicted values of the market price and each of the elements.
[Selected figure] Figure 7

Description

The present invention relates to an information processing device, an information processing method, and an information processing program.

In recent years, new business models have been considered for electric power companies that operate renewable energy generation facilities and power storage facilities.

When new business models are realized, technology is sometimes used to create operational plans for power facilities, taking into account uncertainties such as future electricity market prices and electricity demand (for example, Patent Document 1).

Patent No. 4154373

In the invention described in Patent Document 1, an operational plan is created based on multiple scenarios created using random numbers. However, a large number of scenarios are required to ensure the accuracy of the operational plan, which can result in high calculation costs.

The present invention was made in consideration of these problems, and aims to provide an information processing device that can reduce the calculation costs when creating an operation plan for power equipment.

One invention for achieving the above object is an information processing device comprising: a first acquisition unit that acquires first data including a plurality of past predicted values of the market price of electricity and an element related to the market price and actual values corresponding to the past predicted values; a second acquisition unit that acquires second data obtained by downsampling the first data so as to maintain a first error trend between the actual value and the predicted value of the market price and a second error trend between the actual value and the predicted value of the element; and a scenario creation unit that creates a plurality of scenarios for the future market price and each of the elements based on the second data and the future predicted values of the market price and each of the elements.

The information processing method includes a step of acquiring first data including a plurality of past predicted values of the electricity market price and elements related to the market price and actual values corresponding to the past predicted values, a step of acquiring second data by downsampling the first data so as to maintain a first error trend between the actual value and the predicted value of the market price and a second error trend between the actual value and the predicted value of the element, and a step of creating a plurality of future scenarios for the market price and each of the elements based on the second data and the future predicted values of the market price and each of the elements.

The present invention also provides an information processing program that causes a computer to implement a first acquisition unit that acquires first data including a plurality of past predicted values of the market price of electricity and an element related to the market price and actual values corresponding to the past predicted values, a second acquisition unit that acquires second data obtained by downsampling the first data so as to maintain a first error trend between the actual value and the predicted value of the market price and a second error trend between the actual value and the predicted value of the element, and a scenario creation unit that creates a plurality of scenarios for the future market price and each of the elements based on the second data and the future predicted values of the market price and each of the elements. Other features of the present invention will be made clear by the description in this specification.

The present invention provides an information processing device that can reduce the calculation costs when creating an operation plan for power equipment.

FIG. 1 is a diagram illustrating an information processing system 1 according to an embodiment. 11 is a diagram illustrating data D2 output by the prediction device 2. FIG. 13 is a diagram explaining data D3 stored in the uncertainty element database DB2. FIG. FIG. 2 is a diagram illustrating a hardware configuration of an information processing device 3 according to the embodiment. FIG. 2 is a diagram illustrating functional blocks of an information processing device 3 according to an embodiment. FIG. 11 is a diagram for explaining an error E1 and an error E2. FIG. 13 is a diagram for explaining the tendency of an error E1. FIG. 13 is a diagram illustrating an error E1 and an error E2 after downsampling of data D3. FIG. 2 illustrates an example of distribution of market prices and electricity demand before and after downsampling. FIG. 11 is a diagram for explaining a plurality of scenarios that have been created. FIG. 11 is a diagram for explaining a created operation plan. 1 is a diagram explaining the relationship between expected profits and risk in each of the created management plans. FIG. 10 is a flowchart illustrating a process executed by an information processing device 3. 10 is a flowchart illustrating a process executed by an information processing device 3.

==Embodiment==
<<Information Processing System 1>>
The information processing system 1 in FIG. 1 is a system for creating an operation plan for a power facility 4, taking into consideration uncertainties such as future market prices of electricity and electricity demand.

Figure 1 is a diagram illustrating an information processing system 1 according to this embodiment. The information processing system 1 includes an initial value database DB1, a prediction device 2, an uncertainty element database DB2, a setting value database DB3, and an information processing device 3.

<Initial value database DB1>
The initial value database DB1 stores data D1 relating to explanatory variables to be input to a prediction device 2 (details of which will be described later).

Data D1 is, for example, information on electricity supply and demand, meteorological information such as temperature and solar radiation, and information on the electricity market such as market prices. As data D1, the most recent data (e.g., data from the past month, past three months, etc.) for the period for which the operation plan is to be created (hereinafter referred to as the "planning period") is used.

<Prediction device 2>
The prediction device 2 is a device that predicts uncertainties during a future planning period designated by a user, using the above-mentioned data D1. The uncertainties predicted by the prediction device 2 include, for example, market prices, electricity demand, the amount of power generated by each power facility 4, and electricity trading unit prices.

The prediction device 2 calculates a predicted value of the market price, for example, based on at least the market price and the amount of solar radiation in the data D1. The prediction device 2 also calculates a predicted value of the electricity demand based on at least the electricity demand and the temperature in the data D1.

The predicted market price and electricity demand values calculated by the prediction device 2 are output as data D2 and stored in the uncertainty element database DB2.

Figure 2 is a diagram explaining data D2 output by the prediction device 2. In this example, the market price (Figure 2(a)) and electricity demand (Figure 2(b)) are shown as uncertainties.

In this example, the date (base date) on which the operation plan is created is May 30, 2024, and the planning period is from 0:00 to 23:00 on June 1 of the same year. In this case, the prediction device 2 predicts the uncertainty factors for the planning period (June 1 of the same year) on the base date (May 30 of the same year).

Figure 2 shows the forecast values of the market price and electricity demand on June 1, 2024. Data D2 also includes future forecast values of the market price and future forecast values of electricity demand for each of the multiple divided time periods of a day. In this example, the time period of a day is divided into 24 time periods (0:00 to 23:00) at one-hour intervals.

<Uncertainty Element Database DB2>
The uncertainty element database DB2 stores data D2 output from the prediction device 2, and actual values and past predicted values (data D3) of the uncertainty elements.

Data D3 (corresponding to "first data") is data that includes multiple past forecast values and actual values corresponding to the past forecast values for the market price of electricity (hereinafter sometimes simply referred to as "market price") and uncertainty factors related to the market price (corresponding to "factors").

The uncertainty element database DB2 further stores the equipment status (such as the SOC of the storage battery) from the power equipment 4, power generation results, and transaction results from the business partner 5.

Figure 3 is a diagram explaining data D3 stored in the uncertainty element database DB2. In this embodiment, the market price of electricity (Figure 3(a)) and electricity demand (Figure 3(b)) are considered as uncertainty elements. In the example of Figure 3, data D3 includes 91 days of data from April 1, 2023 to June 30, 2023 for both the market price and electricity demand. Note that this 91-day period may be referred to as the "learning period" below.

The past predicted values in the data D3 are the predicted values in the data D2 that were previously output by the prediction device 2. The actual values in the data D3 are the actual values output by the power facility 4 and the customer 5.

Data D3 includes actual and past forecast values of market prices for each of a number of divided time periods of a day, and actual and past forecast values of electricity demand. In this example, a time period of a day is divided into 24 time periods (0:00 to 23:00) at one-hour intervals.

In the following description, the market price and electricity demand may be collectively referred to as "uncertainty factors." Note that, in this embodiment, the market price and electricity demand are taken into consideration as uncertainty factors, but this example is not limiting. Uncertainty factors are data indicating future events that are necessary when predicting market prices.

In this embodiment, the uncertainty factors include at least the market price. The other uncertainty factors may be related to the market price of electricity, and may be, for example, the amount of solar radiation, the amount of power generated by the power facility, etc., instead of the electricity demand, or may be, for example, the amount of solar radiation, the amount of power generated by the power facility, etc., in addition to the electricity demand.

<Setting value database DB3>
The set value database DB3 stores data D4 including information on the planning period, the trading partner 5, etc. In this embodiment, the planning period is from 0:00 to 23:00 on June 1, 2024 (FIG. 2). The trading partner 5 is, for example, an electricity trading market such as a supply and demand adjustment market or a wholesale electricity market.

<Information processing device 3>
4 is a diagram for explaining the hardware configuration of the information processing device 3 of this embodiment. The information processing device 3 is a computer having a CPU (Central Processing Unit) 300, a memory 301, a communication device 302, a storage device 303, an input device 304, an output device 305, and a recording medium reading device 306.

[CPU 300]
The CPU 300 executes information processing programs stored in the memory 301 and the storage device 303 to realize various functions of the information processing device 3 .

[Memory 301]
The memory 301 is, for example, a RAM (Random-Access Memory) and is used as a temporary storage area for various programs, data, and the like.

[Communication device 302]
The communication device 302 exchanges various programs and data with other computers via the network.

[Storage device 303]
The storage device 303 is a non-transitory (e.g., non-volatile) storage device that stores various data to be executed or processed by the CPU 300 .

[Input device 304]
The input device 304 is a device that accepts commands and data input by a user, and includes an input interface such as a keyboard and a touch sensor that detects a touch position on a touch panel display.

[Output device 305]
The output device 305 is, for example, a display or a printer.

[Recording medium reader 306]
The recording medium reader 306 reads various data such as information processing programs recorded on recording media such as SD cards, DVDs, and CD-ROMs, and stores the data in the storage device 303 .

5 is a diagram showing functional blocks of the information processing device 3. The information processing device 3 includes acquisition units 310 and 311, a scenario creation unit 312, an operation plan creation unit 313, a risk assessment unit 314, and an output unit 315.

[Acquisition unit 310]
The acquiring unit 310 acquires data D2 (FIG. 2) and data D3 (FIG. 3). In this embodiment, the acquiring unit 310 acquires these data from the uncertainty element database DB2.

[Acquisition unit 311]
The acquisition unit 311 acquires data D4 (corresponding to "second data") obtained by downsampling the data D3.

Here, "downsampling" refers to extracting data for some dates from the data for dates included in data D3 (in the example of FIG. 3, the 91 days from April 1, 2023 to June 30, 2023). Downsampling is described in detail below.

When downsampling, the acquisition unit 311 calculates, from the data D3, an error E1 (corresponding to a "first error") between the actual value and the predicted value of the market price, and an error E2 (corresponding to a "second error") between the actual value and the predicted value of the electricity demand. Note that when there is no distinction between the error E1 and the error E2, they may be referred to as error E.

In this embodiment, each of the errors E1 and E2 is a relative error, i.e., the difference between the predicted value and the actual value of the uncertainty element is divided by the actual value, as shown in the following formula.
Relative error = (predicted value - actual value) / actual value

Figure 6 is a diagram explaining error E1 and error E2. In this diagram, d1 to d91 correspond to dates, and respectively correspond to April 1, 2023 to June 30, 2023.

Figure 6(a) shows the market price error E1 on dates and in each time period during the learning period. Furthermore, Figure 6(a) shows the absolute values of the market price error E1 in each time period added up over the learning period as the "total." Hereinafter, "total error" refers to the value obtained by adding up the absolute values of the error over the learning period.

As shown in Figure 6(a), for example, the error E1 at 0:00 on d=1 (April 1, 2023) is 2.0%. The total error E1 at 0:00 is the sum of the absolute values of the errors E1 at d=1 to d=91, which is 170.1%. Note that Figure 6(b) is similar to Figure 6(a) and relates to the error E2 in power demand.

The acquisition unit 311 downsamples the data D3 so that the trend of the error E1 and the trend of the error E2 are maintained. In this embodiment, the "trend of the error E1" refers to the relative frequency of each class of the error E1 in the time period when the sum of the errors E1 is at its maximum. The same applies to the "trend of the error E2."

Furthermore, "the trend of error E1 is maintained" means that the shape of the frequency distribution of error E1 is maintained before and after downsampling. Specifically, this means that after downsampling, the fluctuations in statistics such as the expected value, standard deviation, kurtosis, skewness, etc. of error E1 are suppressed within a specified range. The same applies to error E2. This will be explained in detail below.

The acquisition unit 311 identifies the time period when the sum of the errors E1 is the largest and the time period when the sum of the errors E2 is the largest. In this example, from FIG. 6(a), the time period when the sum of the errors E1 is the largest is 12:00, and the sum is 274.1%. Also, from FIG. 6(b), the time period when the sum of the errors E2 is the largest is 14:00, and the sum is 221.1%.

The time period when the sum of errors E1 is maximum corresponds to the "first time period," and the time period when the sum of errors E2 is maximum corresponds to the "second time period."

Figure 7 is a diagram explaining the trend of errors. Figure 7 is a diagram showing the distribution of the relative frequency of error E1, with the horizontal axis representing the relative error and the vertical axis representing the relative frequency. Here, the relative error on the horizontal axis is the error E1 at 12:00, the time period in which the total of error E1 is the largest.

The relative frequency on the horizontal axis is divided into a set number of classes. The relative frequency on the vertical axis is the standardized value of the number of relative errors belonging to each class.

The solid line is the distribution of relative frequencies obtained from data D3 (Figure 3) that is market price (sometimes referred to as "before downsampling"), and the bar graph is the distribution of relative frequencies obtained from data D4 that is market price (sometimes referred to as "after downsampling") obtained by downsampling data D3.

As can be seen from Figure 7, the relative error peaks near 0% before and after downsampling, and the frequency distributions before downsampling (solid line) and after downsampling (bar graph) are similar to each other. Furthermore, when the sign of the relative error is inverted before and after downsampling, the frequency distribution becomes asymmetric.

In other words, the acquisition unit 311 downsamples the data D3 so that the relative frequency of each class of the error E1 and the error E2 is maintained. At this time, the acquisition unit 311 can downsample using a stratified sampling method.

Figure 8 is a diagram explaining the error E1 and error E2 after downsampling data D3. In this example, 91 days' worth of data contained in data D3 (Figure 3) is downsampled, and 10 days' worth of data is obtained.

For market prices, data for 10 days was obtained: d = 6 (April 6, 2023), d = 11 (the 11th of the same month of the same year), ..., d = 90 (June 29, 2023) (Figure 8 (a)). For electricity demand, data for 10 days was obtained: d = 4 (April 4, 2023), d = 13 (the 13th of the same month of the same year), ..., d = 91 (June 30, 2023) (Figure 8 (b)).

As can be seen from Figure 8, the 10 days of dates obtained by downsampling the market price do not necessarily match the 10 days of dates obtained by downsampling the electricity demand.

However, even with the downsampling method described above, the correlation between the market price and electricity demand is maintained even after downsampling. This will be explained in detail using Figure 9.

Figure 9 shows an example of the distribution of market prices and electricity demand before and after downsampling. In Figure 9, data at time 0:00 is shown. Figure 9 shows frequency distributions (a) and (b) and a scatter plot (c).

Frequency distribution (a) is the frequency distribution of market prices, and frequency distribution (b) is the frequency distribution of electricity demand. In frequency distributions (a) and (b), the solid lines show the data before downsampling, and the dashed lines show the data after downsampling.

As can be seen from frequency distributions (a) and (b), both the market price and electricity demand show distributions that deviate significantly from, for example, a normal distribution.

In scatter plot (c), black circles indicate data before downsampling, and white circles indicate data after downsampling. In scatter plot (a), before downsampling, one piece of data (black circles) is data from the same date. On the other hand, after downsampling, one piece of data (white circles) is not necessarily data from the same date.

After downsampling, when the market price and electricity demand are sorted in ascending order by date, data for dates corresponding to the same rank form a single data point (white circle).

For example, the data for d=6 (Figure 8(a)), which is the earliest date for market price, and the data for d=4 (Figure 8(b)), which is the earliest date for electricity demand, form one data point (white circle).

Furthermore, the data for d=46 (Figure 8(a)), which is the fifth earliest date for market price, and the data for d=44 (Figure 8(b)), which is the same date (fifth earliest) for electricity demand, form one data point (white circle).

As can be seen from Figure 9, the data distribution is similar before and after downsampling. In other words, areas where data (black circles) are dense (sparse) before downsampling also have dense (sparse) data (white circles) before downsampling.

Furthermore, for both the market price and the electricity demand, the frequency distributions (a) and (b) deviate significantly from the normal distribution. However, even in such cases, the correlation between the market price and the electricity demand is taken into account by the downsampling method described above.

[Scenario Creation Unit 312]
The scenario creation unit 312 in Fig. 5 creates multiple scenarios for the future market price and electricity demand based on data D4 obtained by downsampling data D3 (Fig. 3) and the future predicted values of the market price and electricity demand (Fig. 2). The scenarios here include a market price scenario and an electricity demand scenario.

The scenario creation unit 312 creates a scenario by adding an amount corresponding to the relative error of data D4 (Figure 8) for each time period to the future forecast value of the market price (Figure 2 (a)).

The amount according to the relative error is, for example, the absolute error that is expected to be included in the future predicted value based on the relative error. Below, we will explain using a specific example.

Figure 10 is a diagram explaining the multiple scenarios that were created. In this example, ten scenarios (scenario 1 to scenario 10) were created for each of the market price and electricity demand.

Market price scenario 1 is a scenario created by correcting the market price forecast value based on the relative error of d = 6 in Figure 8 (a).

For example, the market price value at 0:00 for scenario 1 (20.7 yen) is the value (20.0 yen + 20.0 yen x 3.4%) obtained by correcting the predicted market price value at 0:00 (20.0 yen) based on the relative error at 0:00 for d=6 (3.4% ).

The market price scenario 1 value at 1:00 (16.5 yen) is the predicted market price value at 1:00 (16.0 yen) corrected based on the relative error at 1:00 for d=6 (3.2%; 16.0 yen + 16.0 yen x 3.2%). The same applies to other time periods for market price scenario 1.

The market price scenario 2 value at 0:00 (19.1 yen) is the predicted market price value at 0:00 (20.0 yen) corrected based on the relative error at 0:00 for d=11 (-4.3%; 20.0 yen - 20.0 yen x 4.3%). The same applies to other market price scenarios and other time periods.

The value at 0:00 for power demand scenario 1 (6.6 [MWh]) is the predicted power demand value at 0:00 (6.4 [MWh]) corrected based on the relative error at 0:00 for d=4 (3.6 [%]) (6.4 [MWh] + 6.4 [MWh] x 3.6 [%]). The same applies to other power demand scenarios and other time periods.

In this embodiment, the relative error is used as the error between the actual value and the past predicted value, but this is not limited to this. Instead of the relative error, the absolute error (the difference between the actual value and the past predicted value) may be used.

[Operation plan creation unit 313]
5 creates an operation plan for the power facility 4 based on the scenario created by the scenario creation unit 312. The operation plan includes, for example, an electricity trading plan, a facility plan, a wheeling plan, and the like.

The operation plan creation unit 313 creates an operation plan, for example, by solving an optimal planning problem. At this time, the created scenario is treated as a fixed variable. The operation plan creation unit 313 may also create an operation plan, for example, by solving a mixed integer linear programming problem.

Figure 11 is a diagram explaining the created operation plan. In this embodiment, the operation plan is a power trading plan. In Figure 11, ten trading plans (plans 1 to 10) are shown. Plans 1 to 10 are created based on scenarios 1 to 10, respectively.

In Figure 11, the electricity trading plan is shown in a bar graph for each plan. In the electricity trading plan, the horizontal axis is time, and the vertical axis is the amount of electricity sold (if positive) or the amount of electricity purchased (if negative).

[Risk Assessment Unit 314]
The risk assessment unit 314 in FIG. 5 performs risk assessment based on statistics relating to the profits that will be obtained when the power facility 4 is operated according to the created operation plan.

Here, the "statistics related to revenue" is, for example, the expected value of revenue (expected revenue). Figure 11 shows, for each plan, the expected revenue when the power facility 4 is operated based on each plan.

In addition, "risk" in risk assessment refers to, for example, the standard deviation of returns, statistics based on the standard deviation, etc. Statistics based on standard deviation refer to, for example, variance, the ratio of standard deviation to expected value, etc. Figure 11 shows, for each plan, the ratio of standard deviation to expected return as the risk when operating based on each plan.

[Output unit 315]
The output section 315 in FIG. 5 outputs the relationship between the expected return and the risk evaluated by the risk evaluation section 314 to the output device 305 .

Figure 12 is a diagram explaining the relationship between expected returns and risk in each of the created investment plans (Figure 11). In Figure 12, the horizontal axis is risk and the vertical axis is expected returns. In this diagram, the further to the right the risk, and the higher the expected returns.

<Processing Executed by Information Processing Device 3>
The process executed by the information processing device 3 will be described with reference to a flowchart. Fig. 13 is a flowchart illustrating the process executed by the information processing device 3. Fig. 14 is a flowchart illustrating the details of downsampling among the processes executed by the information processing device 3.

First, in step S11, the acquisition unit 310 acquires a future predicted value of the market price from among the uncertainty factors. The above-mentioned FIG. 2(a) is an example of a future predicted value of the market price from the data D2.

Next, in step S12, the acquisition unit 310 acquires past predicted and actual market price values. FIG. 3(a) described above is an example of past predicted and actual market price values from data D3.

Next, in step S13, the acquisition unit 311 acquires data obtained by downsampling the past predicted values and actual values of the market price acquired in step S12. The process of step S13 will be described in detail with reference to FIG. 14.

First, in step S131, the acquisition unit 311 calculates the error E1 between the actual value and the predicted value from the past predicted value and actual value of the market price acquired in step S12. The above-mentioned FIG. 6(a) is an example of the calculated error E1.

Next, in step S132, the acquisition unit 311 calculates the sum of the errors E1 for each time period. Figure 6 (a) shows the sum of the errors E1 for each time period.

Next, in step S133, the acquisition unit 311 identifies the time period in which the sum of the errors E1 is maximum. In the example of FIG. 6(a), the time period in which the sum of the errors E1 is maximum is 12:00.

Next, in step S134, the acquisition unit 311 creates a frequency distribution of the error E1. The above-mentioned FIG. 7 is an example of the frequency distribution of the error E1.

Next, in step S135, the acquisition unit 311 downsamples the data D3 of the past predicted and actual market price values acquired in step S12 of FIG. 13.

Figure 8 (a) shows an example of data after downsampling, where the original 91 days of data has been downsampled to obtain 10 days of data.

In S135, the acquisition unit 311 acquires data for some dates so that the trend of the error E1 is maintained. The aforementioned FIG. 7 shows a relative frequency distribution of the error E1 as an example of the trend of the error E1.

In Figure 7, the solid line shows the distribution of relative frequencies of market prices before downsampling, and the bar graph shows the distribution of relative frequencies of market prices after downsampling.

Returning to FIG. 13, steps S11 to S13 are repeated for other uncertain factors of the market price. In this embodiment, the other uncertain factor of the market price is electricity demand.

Next, in step S14, the scenario creation unit 312 creates multiple scenarios for future market prices and electricity demand based on the data acquired in step S13 (Figure 8) and the future forecast values acquired in step S11 (Figure 2). Figure 10 above is an example of the multiple scenarios that have been created.

Next, in step S15, the operation plan creation unit 313 creates an operation plan for the power equipment 4 based on the scenario created in step S14. Figure 11 is an example of the created operation plan.

Next, in step S16, the risk assessment unit 314 performs risk assessment based on statistics regarding the profits to be obtained when the power facility 4 is operated according to the operation plan made in step S15. FIG. 11 shows expected profits as statistics regarding profits, and the standard deviation of expected profits as risk.

Next, in step S17, the output unit 315 outputs the relationship between expected return and risk evaluated by the risk evaluation unit 314 to the output device 305. The aforementioned FIG. 12 is a diagram explaining the relationship between expected return and risk evaluated in step 15.

The information processing device 3 of this embodiment described above makes it possible to reduce the calculation costs when creating an operation plan for the power equipment 4.

==Summary==
The information processing device 3 of the embodiment described above includes an acquisition unit 310 that acquires data D3 including multiple past predicted values of the electricity market price and uncertainty factors related to the market price and actual values corresponding to the past predicted values, an acquisition unit 311 that acquires data D4 obtained by downsampling the data D3 so as to maintain the trend of the error E1 between the actual value and the predicted value of the market price and the trend of the error E2 between the actual value and the predicted value of the uncertainty factors, and a scenario creation unit 312 that creates multiple scenarios for the future market price and each of the elements based on the data D4 and the future predicted values of the market price and each of the elements.

With this configuration, future prediction scenarios are created using the downsampled data D4, making it possible to reduce calculation costs when creating operational plans for power facilities.

In addition, in the information processing device 3, the data D3 includes the actual value and past forecast value of the market price in a first time period among the multiple divided time periods of the day, and the actual value and past forecast value of the element in a second time period. With this configuration, data for all time periods of the multiple divided time periods of the day is not used, so it is possible to further reduce the calculation cost when creating an operation plan for the power equipment.

In addition, in the information processing device 3, the first time period is the time period in which the sum of errors E1 is maximum, and the second time period is the time period in which the sum of errors E2 is maximum. With this configuration, it is possible to create a future scenario that fully takes into account anticipated risks.

In addition, in the information processing device 3, each of the errors E1 and E2 is a relative error. With this configuration, it is possible to create a future scenario that does not deviate significantly from the trend of the actual values of the uncertainty elements and is close to the trend of the actual values.

In addition, in the information processing device 3, the acquisition unit 311 downsamples the data D4 so that the relative frequency of each class of the error E1 and the error E2 is maintained. With this configuration, the correlation between the uncertainty elements is taken into account.

The information processing device 3 further includes an operation plan creation unit 313 that creates an operation plan for the power equipment based on the created scenario. With this configuration, it is possible to create an operation plan while keeping calculation costs down.

The information processing device 3 further includes a risk assessment unit 214 that performs risk assessment based on statistics regarding the profits that can be obtained when the power equipment is operated according to the created operation plan. With this configuration, it is possible to perform risk assessment while suppressing calculation costs.

The information processing method of the embodiment includes the steps of: an information processing device 3 acquiring data D3 including multiple past predicted values of the electricity market price and the uncertainty elements related to the market price and actual values corresponding to the past predicted values; acquiring data D4 by downsampling the data D3 so that the trend of the error E1 between the actual value and the predicted value of the market price and the trend of the error E2 between the actual value and the predicted value of the uncertainty elements are maintained; and creating multiple scenarios for the future market price and each of the elements based on the data D4 and the future predicted values of the market price and each of the elements.

According to this method, future prediction scenarios are created using the downsampled data D4, which makes it possible to reduce the calculation costs when creating operational plans for power facilities.

The information processing program causes the computer to implement an acquisition unit 310 that acquires data D3 including multiple past predicted values of the electricity market price and the uncertainty elements related to the market price and actual values corresponding to the past predicted values, an acquisition unit 311 that acquires data D4 obtained by downsampling the data D3 so as to maintain the trend of the error E1 between the actual value and the predicted value of the market price and the trend of the error E2 between the actual value and the predicted value of the uncertainty elements, and a scenario creation unit 312 that creates multiple scenarios for the future market price and each of the elements based on the data D4 and the future predicted values of the market price and each of the elements.

According to such a program, future prediction scenarios are created using the downsampled data D4, making it possible to reduce calculation costs when creating operational plans for power facilities.

Information Processing System 1
Prediction device 2
Information processing device 3
CPU 300
Memory 301
Communication device 302
Storage device 303
Input device 304
Output device 305
Recording medium reader 306
Acquisition unit 310
Acquisition unit 311
Scenario creation section 312
Operational planning department 313
Risk Assessment Division 314
Output unit 315

Claims (7)

a first acquisition unit that acquires first data including a plurality of past predicted values of a market price of electricity and an element related to the market price, and actual values corresponding to the past predicted values;
a second acquisition unit that acquires second data obtained by downsampling the first data so that a trend of a first error between the actual value and the predicted value of the market price and a trend of a second error between the actual value and the predicted value of the element are maintained;
a scenario creation unit that creates a plurality of scenarios for the future market price and each of the elements based on the second data and future predicted values of the market price and each of the elements,
the first data is the actual value and the past predicted value of the market price in a first time period among a plurality of time periods divided into one day, and the actual value and the past predicted value of the element in a second time period ;
the first time period is a time period in which the sum of the first errors is maximum,
The second time period is a time period in which the sum of the second errors is maximum.
Information processing device. 2. The information processing device according to claim 1,
Each of the first and second errors is a relative error.
Information processing device. 2. The information processing device according to claim 1 ,
The second acquisition unit is
downsampling the first data such that the relative frequency of each class of the first and second errors is maintained;
Information processing device. The information processing device according to any one of claims 1 to 3 ,
Further comprising an operation plan creation unit that creates an operation plan for the power equipment based on the scenario.
Information processing device. 5. The information processing device according to claim 4 ,
A risk assessment unit that performs risk assessment based on statistics regarding profits obtained when the power equipment is operated according to the operation plan.
Information processing device. An information processing device,
acquiring first data including a plurality of past predicted values of a market price of electricity and a factor related to the market price, and actual values corresponding to the past predicted values;
obtaining second data by downsampling the first data so as to maintain a first error trend between the actual value and the forecast value of the market price and a second error trend between the actual value and the forecast value of the element;
generating a plurality of future scenarios of the market price and each of the factors based on the second data and future forecast values of the market price and each of the factors;
the first data is the actual value and the past predicted value of the market price in a first time period among a plurality of time periods divided into one day, and the actual value and the past predicted value of the element in a second time period;
the first time period is a time period in which the sum of the first errors is maximum,
The second time period is a time period in which the sum of the second errors is maximum.
Information processing methods . On the computer,
a first acquisition unit that acquires first data including a plurality of past predicted values of a market price of electricity and an element related to the market price, and actual values corresponding to the past predicted values;
a second acquisition unit that acquires second data obtained by downsampling the first data so that a trend of a first error between the actual value and the predicted value of the market price and a trend of a second error between the actual value and the predicted value of the element are maintained;
a scenario creation unit that creates a plurality of scenarios for the future market price and each of the elements based on the second data and future predicted values of the market price and each of the elements,
the first data is the actual value and the past predicted value of the market price in a first time period among a plurality of time periods divided into one day, and the actual value and the past predicted value of the element in a second time period;
the first time period is a time period in which the sum of the first errors is maximum,
The second time period is a time period in which the sum of the second errors is maximum.
Information processing program .

Images