JP7653052B2 - Power supply planning support device and power supply planning support program - Google Patents

Description

This invention relates to a power supply planning support device and a power supply planning support program, and more specifically to technology that supports the formulation of plans for the continuation or abolition of power supplies, and the establishment or cancellation of power supplies, based on profitability.

A known conventional technology for predicting the contract price of electricity traded in the electricity trading market is a system that stores information on environmental performance, including past weather information, fuel price performance, and other information in association with electricity market prices, and predicts electricity market prices based on environmental forecast information, including weather information, and fuel prices (see Patent Document 1).

JP 2019-046281 A

In order to formulate a medium- to long-term power supply plan, it is necessary to evaluate the profitability of each power source and obtain information on the change in profits when each existing power source is maintained or abolished, and when a new power source is newly established. However, the system described in Patent Document 1 cannot predict the change in profits when each existing power source is maintained or abolished, and when a new power source is newly established.

The present invention aims to provide a power supply plan formulation support device and a power supply plan formulation support program that can evaluate the profitability of each power source and provide information on changes in profits when existing power sources are maintained or abolished and when new power sources are newly established.

In order to solve the above problem, the invention of claim 1 is a power supply plan formulation support device characterized by having: means for calculating the energy volume revenue of the power supply by multiplying the amount of power generated by the power supply by the market energy price; means for calculating the capacity revenue of the power supply by multiplying the power generation capacity of the power supply by the market supply capacity price; means for calculating the revenue of the power supply by subtracting the sum of the fuel costs and fixed costs of the power supply from the sum of the energy volume revenue and the capacity revenue ; and means for determining that the existing power supply should continue to exist if the revenue is equal to or greater than a predetermined revenue threshold, and for determining that the existing power supply should be abolished if the revenue is less than the predetermined revenue threshold .

The invention of claim 2 is a power supply plan formulation support device characterized by having: means for calculating the energy revenue of the power supply by multiplying the amount of power generated by the power supply by the market energy price; means for calculating the capacity revenue of the power supply by multiplying the power generation capacity of the power supply by the market supply capacity price; means for calculating the revenue of the power supply by subtracting the sum of the fuel costs and fixed costs of the power supply from the sum of the energy revenue and the capacity revenue; means for calculating the rate of profit of the new power supply using the revenue and the initial investment amount of the power supply; and means for determining that the new power supply should be established if the rate of profit is equal to or greater than a predetermined rate of profit threshold, and for determining that the new power supply should be discontinued if the rate of profit is less than the predetermined rate of profit threshold.

The invention of claim 3 is characterized in that, in the power supply plan formulation support device according to claim 1 or 2, it has a means for estimating the electricity price in the market by using a prediction model constructed by machine learning with an electricity price factor that fluctuates the electricity price as an input and an electricity price as an output.

The invention of claim 4 is characterized in that, in the power supply plan formulation support device described in claim 2, it has a means for estimating the profitability of the new power source by using a prediction model constructed by machine learning, with profitability factors having a significant effect on the profitability as input and the profitability as output.

The invention of claim 5 is a power source plan formulation support program that causes a computer to function as: means for calculating the energy revenue of a power source by multiplying the amount of power generated by the power source by the market energy price; means for calculating the capacity revenue of the power source by multiplying the power generation capacity of the power source by the market supply capacity price; means for calculating the revenue of the power source by subtracting the sum of the fuel costs and fixed costs of the power source from the sum of the energy revenue and the capacity revenue; and means for determining that the existing power source should continue to exist if the revenue is equal to or greater than a predetermined revenue threshold, and for determining that the existing power source should be abolished if the revenue is less than the predetermined revenue threshold.

The invention of claim 6 is a power supply plan formulation support program that causes a computer to function as: means for calculating the energy revenue of a power supply by multiplying the amount of power generated by the power supply by the market energy price; means for calculating the capacity revenue of the power supply by multiplying the power generation capacity of the power supply by the market supply capacity price; means for calculating the revenue of the power supply by subtracting the sum of the fuel costs and fixed costs of the power supply from the sum of the energy revenue and the capacity revenue; means for calculating the rate of profit of a new power supply using the revenue and the initial investment amount of the power supply ; and means for determining that the new power supply should be established if the rate of profit is equal to or greater than a predetermined rate of profit threshold, and for determining that the new power supply should be discontinued if the rate of profit is less than the predetermined rate of profit threshold.

The invention of claim 7 is characterized in that, in the power supply plan formulation support program of claim 5 or 6, a computer is caused to function as a means for estimating the electricity price in the market by using a prediction model constructed by machine learning with electricity price factors that fluctuate the electricity price as input and the electricity price as output.

The invention of claim 8 is characterized in that, in the power supply plan formulation support program of claim 6, a computer is caused to function as a means for estimating the profitability of the new power source by using a prediction model constructed by machine learning with profitability factors having a significant effect on the profitability as input and with the profitability as output.

According to the inventions of claims 1 and 5, the decision to continue or discontinue an existing power source is made based on the revenue calculated by subtracting the sum of the fuel costs and fixed costs of the power source from the sum of the electricity volume revenue and the capacity revenue . Therefore, it is possible to evaluate the profitability of each power source using the revenue as an evaluation index and provide information regarding changes in revenue when each existing power source is continued or discontinued .

According to the inventions of claims 2 and 6, the decision to install or cancel a new power source is made based on the rate of return calculated using the revenue obtained by subtracting the total of the fuel costs and fixed costs of the power source from the total of the electricity volume revenue and the capacity revenue and the initial investment amount of the power source. Therefore, it is possible to evaluate the profitability of each power source using the rate of return as an evaluation index and provide information on changes in profits when a new power source is installed .

According to the inventions of claims 3 and 7, the market price of electricity is estimated using a prediction model constructed by machine learning with the electricity price factors as input and the electricity price as output, so it is possible to estimate a more realistic price of electricity that reflects actual information on past electricity prices, which in turn makes it possible to improve the reliability of the technology as a support for the formulation of power supply plans.

According to the inventions of claims 4 and 8, the profitability of a power source is estimated using a prediction model constructed by machine learning with profitability factors as input and profitability as output, making it possible to estimate a more realistic profitability that reflects actual information on past profitability, and ultimately improving the reliability of the technology as a support for the formulation of power source plans.

1 is a functional block diagram showing a schematic configuration of a power supply plan development supporting device according to an embodiment of the present invention; 2 is a diagram showing an example of a processing result of the power supply plan development supporting apparatus of FIG. 1 . FIG. 2 is a conceptual diagram showing an example of the configuration of a performance database.

The present invention will be described below based on the illustrated embodiment.

In this invention, the determination of whether to continue or discontinue existing power sources and whether to establish or discontinue new power sources (i.e., power sources scheduled for establishment or power sources with proposed establishment plans, etc.) in a specified cycle is called a "power source plan," and each cycle when formulating a power source plan is called an "evaluation period." One cycle when formulating a power source plan, i.e., the evaluation period, is not limited to a specific length of time, but it can be set to, for example, one year, six months, or three months.

Power supply plans are basically plans for the future that are formulated with reference to past data and performance, and the length of the planning period covered by the power supply plan is not limited to a specific period, but may be set to, for example, 10, 20, or 30 years. The length of the planning period covered by the power supply plan may also be set taking into account the length of the evaluation period when calculating the rate of return of a new power source by the expected rate of return calculation task 158 described below.

The power source refers to a power generation facility, and includes, for example, nuclear power plants, gas/coal/oil-fired power plants, hydroelectric power plants, and renewable energy power plants (specifically, for example, solar power plants, wind power plants, biomass power plants, geothermal power plants, pumped storage power plants, etc.).

Power supply plans are formulated for each designated region, taking into consideration, for example, the extent/consolidation of the power transmission and distribution network and the area to which electricity is expected to be supplied by the power source. Each area range when formulating a power supply plan is called a "planning area." Note that planning areas are not limited to designated areas within Japan, and may be designated areas in a foreign country.

Figure 1 is a functional block diagram showing the schematic configuration of a power supply plan development support device 1 according to an embodiment of the present invention. Figure 2 is a diagram showing an example of the processing result of the power supply plan development support device 1. In the example shown in Figure 2, for a certain planning area, there are power plants A, B, and C as existing power sources, and power plants α and β as new (newly constructed) power sources. In Figure 2, the length of each evaluation period (i.e., one cycle) is one year, and the evaluation period at the beginning of the planning period covered by the power supply plan is "20XX" and the evaluation period at the end is "20YY". Power supply plans are developed for each planning area, and the following describes a case where processing is performed by the power supply plan development support device 1 for the above-mentioned certain planning area.

The power supply plan formulation support device 1 is a mechanism for supporting the formulation of a power supply plan by evaluating the profitability of each power source and creating information regarding changes in profits when existing power sources are maintained or abolished and new power sources are newly established, and is mainly equipped with an input unit 11, a display unit 12, a storage unit 13, a memory 14, a main task 15, a communication unit 17, and a central processing unit 18 that controls these.

The power supply plan development support device 1 is configured, for example, by installing and executing an application (power supply plan development support program 19) for performing various processes related to supporting the development of a power supply plan and a control program for the entire device on various computers, etc.

The input unit 11 is an interface that has the function of receiving commands from a user and inputting them into the power supply plan formulation support device 1, and is configured, for example, with a keyboard and mouse.

The display unit 12 has functions such as displaying the contents input via the input unit 11 and displaying the processing results of the power supply plan formulation support device 1 (see FIG. 2), and is configured, for example, by a liquid crystal display.

The storage unit 13 is a storage area/storage device that has the function of storing various information, programs, data, etc., and is composed of, for example, a hard disk (HDD: short for Hard Disk Drive).

The memory unit 13 stores the control program for the entire power supply planning support device 1 and the power supply planning support program 19, as well as a power supply database 131 and a performance database 132.

The power source database 131 is a database in which information about each power source is recorded and stored, and is referenced as necessary to obtain information required about the power source when the power source plan development support device 1 performs various processes related to supporting the development of a power source plan.

The power source database 131 specifically records and stores various information about each power source that is required when the power source plan formulation support device 1 performs various processes related to supporting the formulation of a power source plan, such as the planned area to supply power, power generation capacity, fuel type, thermal efficiency, and fixed costs for each power source. In addition to information about each existing power source, the power source database 131 also records and stores information about each new power source, and for new power sources, the initial investment amount (note that this is the initial investment amount required when calculating the rate of return of a new power source by the expected return calculation task 158 described below, in other words, the initial business cost) is also recorded and stored.

The performance database 132 will be described later.

Either or both of the power supply database 131 and the performance database 132 may be stored in an external storage device such as a server. In this case, the power supply plan formulation support device 1 may access the external storage device via the communication unit 17 to acquire various data and information, or may be provided with a connection interface for transmitting/receiving/inputting/outputting signals such as data and control commands to/from the external storage device via various signal lines, and may access the external storage device via the connection interface to acquire various data and information.

The memory 14 is a storage area/storage device that functions as a working area for temporarily storing data and information generated when the central processing unit 18 executes calculations related to supporting the formulation of power supply plans, and is composed of, for example, RAM (short for Random Access Memory).

The main task 15 is a group of tasks for executing various processes related to supporting the formulation of power plans, which are realized by executing the power plan formulation support program 19 stored in the memory unit 13.

The communication unit 17 is a communication interface that has the function of transmitting/receiving/inputting/outputting signals and information transmitted via various wireless/wired communication networks, including, for example, a LAN (short for Local Area Network) and a WAN (short for Wide Area Network).

The central processing unit 18 has the function of coordinating and controlling each component of the power supply plan formulation support device 1, and is configured to include, for example, a central processing unit (CPU: abbreviation for Central Processing Unit), and realizes each function according to the control program stored in the memory unit 13 and the power supply plan formulation support program 19.

The power supply plan formulation support device 1 has a means for calculating the power volume income of the power supply by multiplying the amount of power generated by the power supply by the market price of the power supply (power volume income calculation task 153), a means for calculating the capacity income of the power supply by multiplying the power generation capacity of the power supply by the market price of the supply capacity (capacity income estimation task 154), a means for calculating the revenue of the power supply by subtracting the sum of the fuel cost and fixed cost of the power supply from the sum of the power volume income and the capacity income (revenue calculation task 157), a means for calculating the rate of return of the power supply using the revenue and the initial investment amount of the power supply (expected rate of return calculation task 158), and a means for determining the handling of the power supply based on the revenue and the rate of return (determination task 159).

To execute various processes related to supporting the formulation of power supply plans, the power supply plan formulation support program 19 stored in the memory unit 13 is executed, thereby forming the main task 15. The main task 15 includes an energy amount estimation task 151, a price estimation task 152, an energy amount revenue calculation task 153, a capacity revenue estimation task 154, a fuel cost estimation task 155, a fixed cost estimation task 156, a revenue calculation task 157, an expected rate of return calculation task 158, a judgment task 159, and a learning task 160.

The power amount estimation task 151 is a task that estimates the amount of power generated by each power source [MWh] based on the power demand [MW] for each evaluation period. Specifically, the power demand is estimated as, for example, the hourly power demand.

Specifically, the power consumption estimation task 151 first estimates future power demand (called "base power demand") for each evaluation period based on the future population and industrial location conditions predicted in consideration of future development plans, etc., according to a regression equation that is calculated in advance based on the relationship between power demand and past and present population and industrial location conditions, etc.

The power amount estimation task 151 then, if necessary, corrects the base power demand in accordance with any factors that cause the base power demand to fluctuate, and estimates the final power demand. In other words, if there are factors that reduce power demand in each evaluation period, the base power demand is revised downward in accordance with the degree of impact of that factor on the power demand. Also, if there are factors that increase power demand in each evaluation period, the base power demand is revised upward in accordance with the degree of impact of that factor on the power demand.

Factors that may affect electricity demand include, for example, changes in natural conditions such as climate and temperature, changes in industrial structure, and economic trends.

In addition, data on future power demand that has already been estimated may be used, or the current power demand may be used as the future power demand assuming that the current situation will continue. Here, "assuming" in the explanation of this invention includes setting a value to be estimated in the power supply plan development support device 1, a value that has already been estimated, or a value that remains the current value.

The power amount estimation task 151 further calculates the amount of power generated by each power source in the target planning area of the power source plan that is required according to the power demand for each evaluation period based on the power demand for each evaluation period estimated above. At this time, the constraints are that the amount of power required to cover the power demand is supplied by the sum of the amounts of power generated by each power source supplying power to the target planning area of the power source plan and does not exceed the power generation capability (power generation capacity) of each power source, and it may also be possible to take into account, for example, plans for long-term maintenance work on the power sources.

The power amount estimation task 151 may also take into account the situation of supplying/selling power to other areas that are not the planned area targeted by the power source plan (in other words, the planned area in which the power source under consideration is located), and may add the amount of generated power to be supplied to other areas. In this case, the transmission constraints of the interconnection/transmission system between the planned area and other areas (e.g., the transmission length, voltage, and capacity between areas) may be taken into account.

Then, the power amount estimation task 151 outputs the amount of power generated by each power source for each evaluation period.

The price estimation task 152 is a task that estimates the market price of electricity (specifically, for example, the market price for trading electricity in the wholesale electricity market) [yen/MWh] for each evaluation period.

Here, the unit price of the amount of energy common to all planning areas, ignoring the transmission losses and capacity between planning areas, is called the "system price," while the unit price of the amount of energy that takes into account the transmission losses and capacity between planning areas is called the "energy price." Energy prices usually vary by area due to factors such as uneven distribution of electricity demand and differences in power source composition between planning areas, as well as constraints on the transmission system between planning areas (specifically, transmission losses, transmission capacity, etc.).

Specifically, the price estimation task 152 estimates an energy price (called the "base energy price") for each evaluation period according to a regression equation (i.e., a trend regression equation or trend estimation equation in the trend estimation method) that is calculated in advance based on the market price trend in energy transactions in the wholesale electricity market from the past to the present.

The price estimation task 152 then, if necessary, corrects the base energy price in accordance with any factors that may cause the base energy price to fluctuate, and estimates the final energy price. In other words, if there are factors that push down the energy price in each evaluation period, the base energy price is revised downward in accordance with the degree of impact of those factors on the energy price. Also, if there are factors that push up the energy price in each evaluation period, the base energy price is revised upward in accordance with the degree of impact of those factors on the energy price.

Factors that may affect electricity prices include, for example, the following:
1) Electricity demand A) If the electricity demand (i.e., the total electricity demand by planning area) is large, the system price will be high.

2) Power source information by area a) The larger the nuclear power generation capacity, the cheaper the system price.
C) If the coal-fired power generation capacity is large, the system price will be high.
E) If the renewable energy generating capacity is large, the system price will be lower.
E) If the thermal efficiency is high, the amount of power generated by the power source will be large.

3) Fuel prices by planning area F) If the price of fuel for power generation by planning area increases, the system price will become higher.

4) Information on the power transmission system K) If the transmission length is long, the difference between the system price and the price of electricity per planning area will become larger (this is due to the large transmission loss between planning areas).
h) When the transmission voltage is high, the difference between the system price and the electricity price for each planning area becomes smaller (this is due to smaller transmission losses between planning areas).
i) If the transmission capacity between planning areas is small, the difference between the system price and the electricity price for each planning area will be large (this is because the amount of electricity transmitted will reach the transmission capacity and economical power plants will no longer be able to operate).

5) Electricity demand and power source information J) If there is a large difference in the uneven distribution of electricity demand by planning area or in the power source information by planning area (specifically, for example, power generation capacity, fuel type, thermal efficiency, fixed costs, etc.), the difference between the system price and the electricity price by planning area will become large (Note that this is due to the fact that an increase in the amount of electricity transmitted between planning areas increases transmission losses, and the amount of electricity transmitted reaches the transmission capacity, making it impossible to operate economical power plants).

The price estimation task 152 may also estimate the energy price for each evaluation period according to a multiple regression equation with the energy price factors as explanatory variables and the energy price as the objective variable. In this case, the energy price is estimated according to a multiple regression equation that is determined in advance using a set of combined data of the energy price factors and the actual energy price at multiple points in time from the past to the present for the planning area targeted by the power source plan.

The power price factors, which are explanatory variables in the multiple regression equation, are not limited to specific items, but are appropriately selected and set, taking into consideration whether or not the impact on the power price is significant. Examples of power price factors, which are explanatory variables in the multiple regression equation, include the fluctuation factors of power prices listed as A to J above, such as power demand, nuclear power generation capacity, coal-fired power generation capacity, renewable energy power generation capacity, thermal efficiency, price of fuel for power generation, transmission length, transmission voltage, transmission capacity between planning areas, power demand by planning area, and power source information by planning area (in other words, the contents of the power source configuration).

The price estimation task 152 may alternatively estimate the electricity price for each evaluation period according to a prediction model that uses electricity price factors as input and electricity prices as output. In this case, the electricity price is estimated according to a prediction model (learning model) that is constructed in advance by machine learning using a set of combined data of electricity price factors and actual electricity prices at multiple points in time from the past to the present for the planning area targeted by the power source plan.

As a prediction model (learning model), a machine learning method that is particularly suitable for predicting continuous values may be used. Specifically, for example, an extreme learning machine (ELM: short for Extreme Learning Machine), a deep neural network (DNN: short for Deep Neural Network), or a support vector machine (SVM: short for Support Vector Machine) may be used.

The power price factors that are inputs to the prediction model constructed by machine learning are not limited to specific items, but are appropriately selected and set, taking into consideration whether or not the impact on the power price is significant. Examples of power price factors that are inputs to the prediction model include the fluctuating factors of power prices listed above as A to J, such as power demand, nuclear power generation capacity, coal-fired power generation capacity, renewable energy power generation capacity, thermal efficiency, price of fuel for power generation, transmission length, transmission voltage, transmission capacity between planning areas, power demand by planning area, and power source information by planning area (in other words, the contents of the power source configuration).

In addition, data on future electricity prices that have already been estimated may be used, or the current electricity price may be used as the future electricity price assuming that the current situation will remain unchanged.

Then, the price estimation task 152 outputs the electricity price for each evaluation period.

The power revenue calculation task 153 is a task that calculates the power revenue (in yen) of each power source for each evaluation period.

Specifically, the energy revenue calculation task 153 receives the amount of energy generated by each power source for each evaluation period output from the energy estimation task 151, and receives the energy price for each evaluation period output from the price estimation task 152, and multiplies the amount of energy generated by each of the power sources by the energy price for each evaluation period to calculate the energy revenue of each power source (which is the integral value (in other words, the accumulated value) for the entire evaluation period).

Then, the energy income calculation task 153 outputs the energy income (Re in Figure 2) of each power source for each evaluation period.

In addition, "by evaluation period" means that, as in the example shown in Figure 2, for existing power sources (power plants A, B, and C in the figure), the evaluation period at the beginning of the planning period covered by the power source plan (20XX in the figure) is included as the subject of processing, and for new power sources (power plants α and β in the figure), each evaluation period between the beginning and end of the planning period covered by the power source plan (20YY in the figure) is included as the subject of processing.

In the example shown in FIG. 2, only the evaluation period at the beginning of the planning period covered by the power supply plan (20XX in the figure) is displayed for existing power supplies, but it is also possible to display evaluation periods between the beginning and end of the planning period covered by the power supply plan (20YY in the figure).

In the example shown in FIG. 2, for new (newly installed) power sources, only the evaluation period at the start of the planning period covered by the power source plan (in the figure, 20XX) and the evaluation period at the end of the planning period covered by the power source plan (in the figure, 20YY) are displayed, but it is also possible to display evaluation periods between the evaluation period at the start and the evaluation period at the end of the planning period covered by the power source plan.

The capacity revenue estimation task 154 is a task that estimates the capacity revenue (yen) of each power source for each evaluation period.

Capacity revenue is revenue obtained by trading "power generation capacity" (specifically, supply capacity [MW] equivalent to the amount of electricity (demand) expected to be required in the future). In Japan, "power generation capacity" is traded in a capacity market where the Organization for Cross-regional Coordination of Transmission Operators (OCCO) serves as the market administrator.

Specifically, the capacity revenue estimation task 154 receives the amount of generated power from each power source for each evaluation period output from the power amount estimation task 151, and calculates the power generation reserve [MW] of each power source for each evaluation period, with the constraint that the power generation capacity (power generation capacity) of each power source is not exceeded.

The capacity revenue estimation task 154 also estimates the supply capacity price in the market [yen/MW] (called the "base supply capacity price") for each evaluation period, for example, according to a regression equation (i.e., the trend regression equation or trend estimation equation in the trend estimation method) that is determined in advance based on the market price trend in future supply capacity transactions in the capacity market from the past to the present.

The capacity revenue estimation task 154 then, if necessary, corrects the base supply capacity price in accordance with any factors that may cause the base supply capacity price to fluctuate, and estimates the final supply capacity price. In other words, if there are factors that push down the supply capacity price in each evaluation period, the base supply capacity price is revised downward in accordance with the degree of impact of those factors on the supply capacity price. Also, if there are factors that push up the supply capacity price in each evaluation period, the base supply capacity price is revised upward in accordance with the degree of impact of those factors on the supply capacity price.

The capacity revenue estimation task 154 may take into account "Net CONE" (short for Net Cost Of New Entry), which is an index price in the capacity market, when estimating the base supply capacity price or as a variable factor of the supply capacity price. Net CONE is the level at which a new power source can recover fixed costs (e.g., fixed costs recorded over multiple years, mainly construction costs), and specifically, it is the amount obtained by subtracting the profits obtained in the wholesale electricity market, etc. (i.e., the cash flow after deducting variable costs such as fuel costs from wholesale electricity revenue) from the fixed costs per year of a typical newly constructed power source.

In addition, data on future supply capacity prices that have already been estimated may be used, or the current supply capacity price may be used as the future supply capacity price assuming that the current situation will remain unchanged.

Then, the capacity revenue estimation task 154 calculates the capacity revenue of each power source (which is the integral value (in other words, the accumulated value) for each evaluation period) by multiplying the power generation reserve of each power source for each evaluation period calculated above by the supply capacity price assumed above.

The capacity revenue estimation task 154 may also estimate the capacity revenue of each power source for each evaluation period according to a multiple regression equation with the capacity revenue factors as explanatory variables and the capacity revenue as the objective variable. In this case, the capacity revenue is estimated according to a multiple regression equation that is determined in advance using a set of combined data of the capacity revenue factors and the actual capacity revenue at multiple points in time from the past to the present for the planning area targeted by the power source plan.

The capacity revenue factors, which are explanatory variables in the multiple regression equation, are not limited to specific items, but are appropriately selected and set, taking into consideration whether or not their impact on capacity revenue is significant. Examples of capacity revenue factors, which are explanatory variables in the multiple regression equation, include the power generation capacity (power generation capacity), generated energy volume, and power generation surplus of each power source, as well as the market price in the trading of supply capacity in the capacity market, and "Net CONE", which is the index price in the capacity market.

Alternatively, the capacity revenue estimation task 154 may estimate the capacity revenue of each power source for each evaluation period according to a prediction model that uses capacity revenue factors as input and capacity revenue as output. In this case, the capacity revenue is estimated according to a prediction model (learning model) that is constructed in advance by machine learning using a set of combined data on capacity revenue factors and actual capacity revenue at multiple points in time from the past to the present for the planning area that is the target of the power source plan.

As a predictive model (learning model), a machine learning method that is particularly suitable for predicting continuous values may be used. Specifically, for example, an extreme learning machine (ELM), a deep neural network (DNN), a support vector machine (SVM), etc. may be used.

The capacity revenue factors that are inputs to the prediction model constructed by machine learning are not limited to specific items, but are appropriately selected and set, taking into consideration whether or not the impact on capacity revenue is significant. Examples of capacity revenue factors that are inputs to the prediction model include the power generation capacity (power generation capacity), generated energy amount, and power generation surplus of each power source, as well as the market price for trading supply capacity in the capacity market, and "Net CONE", which is the index price in the capacity market.

Then, the capacity revenue estimation task 154 outputs the capacity revenue (Rc in Figure 2) of each power source for each evaluation period.

The fuel cost estimation task 155 is a task that estimates the fuel cost (in yen) for each power source for each evaluation period, based on the type, specifications, and characteristics of each power source.

Specifically, the fuel cost estimation task 155 estimates the price (unit price; called the "base fuel price") of the power generation fuel for each evaluation period, for example, for each fuel of each power source, according to a regression equation (i.e., the trend regression equation or trend estimation equation in the trend estimation method) that is calculated in advance based on the trend in the price (unit price) of the power generation fuel from the past to the present.

The fuel cost estimation task 155 then, if necessary, corrects the base fuel price in accordance with any factors that may cause the base fuel price to fluctuate, and estimates the final price of fuel for power generation. In other words, if there are factors that push down the price of fuel for power generation in each evaluation period, the base fuel price is revised downward in accordance with the degree of impact of those factors on the price of fuel for power generation. Also, if there are factors that push up the price of fuel for power generation in each evaluation period, the base fuel price is revised upward in accordance with the degree of impact of those factors on the price of fuel for power generation.

Factors that may affect the price of fuel for power generation include, for example, changes in the prices of fossil fuels such as gas, coal, and oil, and changes in the prices of biofuels such as wood biomass.

In addition, data on future prices of fuel for power generation that have already been estimated may be used, or the current price of fuel for power generation may be used as the future price assuming that the current situation will remain unchanged.

Then, the fuel cost estimation task 155 calculates the fuel cost of each power source (which is the integral value (in other words, the accumulated value) for each evaluation period) for each evaluation period based on the price (unit price) of the fuel for power generation for each evaluation period estimated above and the amount of power generated by each power source for each evaluation period output from the power amount estimation task 151.

Fuel cost estimation task 155 may also estimate the fuel cost of each power source for each evaluation period according to a multiple regression equation with fuel cost factors as explanatory variables and fuel costs as the objective variable. In this case, fuel costs are estimated according to a multiple regression equation that is determined in advance using a set of combined data on fuel cost factors and actual fuel costs at multiple points in time from the past to the present for the planning area targeted by the power source plan.

The fuel cost factors, which are explanatory variables in the multiple regression equation, are not limited to specific items, but are appropriately selected and set, taking into consideration whether their impact on fuel costs is significant or not. Examples of fuel cost factors, which are explanatory variables in the multiple regression equation, include the amount of power generated by each power source, as well as the prices of fossil fuels such as gas, coal, and oil, and the prices of biofuels such as woody biomass.

The fuel cost estimation task 155 may alternatively estimate the fuel cost of each power source for each evaluation period according to a prediction model that uses fuel cost factors as input and fuel costs as output. In this case, fuel costs are estimated according to a prediction model (learning model) that is constructed in advance by machine learning using a set of combined data on fuel cost factors and actual fuel costs at multiple points in time from the past to the present for the planning area that is the target of the power source plan.

As a predictive model (learning model), a machine learning method that is particularly suitable for predicting continuous values may be used. Specifically, for example, an extreme learning machine (ELM), a deep neural network (DNN), a support vector machine (SVM), etc. may be used.

The fuel cost factors that are inputs to the prediction model constructed by machine learning are not limited to specific items, but are appropriately selected and set, taking into consideration whether or not their impact on fuel costs is significant. Examples of fuel cost factors that are inputs to the prediction model include the amount of power generated by each power source, as well as the prices of fossil fuels such as gas, coal, and oil, and the prices of biofuels such as woody biomass.

Then, the fuel cost estimation task 155 outputs the fuel cost (Cf in Figure 2) for each power source for each evaluation period.

The fixed cost estimation task 156 is a task that estimates the fixed cost (in yen) of each power source for each evaluation period.

Specifically, the fixed cost estimation task 156 estimates the fixed costs to be recorded for each evaluation period, taking into account the type, specifications, and characteristics of each power source. Note that for fixed costs such as construction costs, the total amount is discounted to the amount per evaluation period. In addition, costs required for operating and maintaining the operation of the power source may be estimated as fixed costs as appropriate.

Then, the fixed cost estimation task 156 outputs the fixed cost (Cc in Figure 2) of each power source for each evaluation period.

The revenue calculation task 157 is a task that calculates the revenue (in yen) of each power source for each evaluation period.

The revenue calculation task 157 specifically receives as input the energy revenue of each power source for each evaluation period output from the energy revenue calculation task 153, the capacity revenue of each power source for each evaluation period output from the capacity revenue assumption task 154, the fuel cost of each power source for each evaluation period output from the fuel cost assumption task 155, and the fixed cost of each power source for each evaluation period output from the fixed cost assumption task 156, and calculates the revenue of each power source (which is the integral value (in other words, the accumulated value) for the entire evaluation period) by subtracting the sum of the fuel cost and the fixed cost from the sum of the energy revenue and the capacity revenue of each power source for each evaluation period.

Then, the revenue calculation task 157 outputs the revenue (R in Figure 2) of each power source for each evaluation period.

Expected profitability calculation task 158 is a task that calculates the profitability of each new power source.

Specifically, the expected return calculation task 158 receives the return of each new power source for each evaluation period output from the return calculation task 157, and calculates the return (also called the "expected return") for each new power source using the return for each evaluation period of the planning period covered by the power source plan (in the example shown in FIG. 2, the return for each year between 20XX and 20YY) and the initial investment amount for that power source.

The rate of return is calculated by setting the length of the evaluation period, taking into account the life cycle of the power source, etc. The length of the planning period covered by the power source plan may be set by taking into account the length of the evaluation period when calculating the rate of return.

Specifically, the expected rate of return calculation task 158 may calculate an index (all expressed in percentages) also known as the internal rate of return (IRR), investment return, or project IRR (PIRR).

Then, the expected return calculation task 158 outputs the return of each new power source (ER in Figure 2).

The judgment task 159 is a task that judges whether each existing power source should continue or be discontinued, and judges whether each new power source should be newly installed or discontinued. Here, the continued existence or discontinuation of a power source, as well as the establishment or discontinuation of a power source, are also referred to as "power source handling."

Specifically, the judgment task 159 receives the revenue for each evaluation period of each existing power source output from the revenue calculation task 157, and judges the power source to continue if the revenue is equal to or greater than a predetermined revenue threshold, and judges the power source to be discontinued if the revenue is less than the predetermined revenue threshold.

The above-mentioned predetermined revenue threshold is not limited to a specific value, but may be set to 0, for example. In this case, if the revenue is not in the red, the power source in question is determined to continue operating, and if the revenue is in the red, the power source in question is determined to be discontinued.

The judgment task 159 also receives the profitability of each new power source output from the expected profitability calculation task 158, and judges it to be a new installation if the profitability is equal to or greater than a predetermined profitability threshold, and judges it to be canceled if the profitability is less than the predetermined profitability threshold.

The above-mentioned predetermined return threshold is not limited to a specific value, and can be set to an appropriate value taking into consideration the type of indicator used as the return rate, but for example, if the internal rate of return (IRR) is used, it may be set to around 8%.

Then, the judgment task 159 outputs the judgment result (J in FIG. 2) of whether each existing power source should be maintained or discontinued, and also outputs the judgment result (J in FIG. 2) of whether each new power source should be established or discontinued.

The results of the determination of whether to establish or cancel each new power source output from the determination task 159 can also be used as a judgment indicator for determining whether to invest in a power source developed by an independent power producer, for example.

Learning task 160 is a task that predicts the profitability of each new power source according to a prediction model that uses profitability factors as input and profitability as output, and determines whether to install or discontinue the power source based on the profitability.

Specifically, the learning task 160 predicts the rate of return according to a prediction model (learning model) constructed by machine learning using a set of combined data of the rate of return factors and the rate of return at multiple points in time from the past to the present for the planning area targeted by the power source plan. The rate of return predicted by the learning task 160 is set to the same type of rate of return as the rate of return calculated in the rate of return calculation process by the expected rate of return calculation task 158.

As a predictive model (learning model), a machine learning method that is particularly suitable for predicting continuous values may be used. Specifically, for example, an extreme learning machine (ELM), a deep neural network (DNN), a support vector machine (SVM), etc. may be used.

The profitability factors that are inputs to the prediction model constructed by machine learning are not limited to specific items, but are appropriately selected and set, for example, by considering whether the impact on the profitability is significant. As profitability factors that are inputs to the prediction model, for example, the planning area and type of power generation of each power source, as well as the generated power amount, power revenue, capacity revenue, fuel cost, fixed cost, and revenue of each power source output from the above tasks 151 to 157, and some or all of the power price in the planning area in which the power source is located are appropriately selected and set, for example, by considering whether the impact on the profitability is significant. In addition, items other than the items output from the above tasks 151 to 157 may be set as profitability factors.

The learning task 160 is configured to execute a learning process for predicting the rate of return on the actual rate of return data recorded and accumulated in the actual database 132 stored in the memory unit 13, construct a prediction model (learning model) based on the actual rate of return data, and obtain an accurate prediction result of the rate of return when specified information and conditions are input.

The performance database 132 is a database in which performance information on the profitability of past power source development is recorded and accumulated, and data on combination data of profitability factors for each power plant for power source development carried out in the past, the profitability calculated based on the combination data, and the actual profitability are associated with each other as past performance information (see Figure 3).

For profitability factors whose values may vary over time (in other words, items whose values may differ depending on the evaluation period; specifically, for example, the amount of generated energy, energy revenue, capacity revenue, fuel costs, fixed costs, and revenue of each power source, as well as the energy price in the planning area in which the power source is located), data for each evaluation period from the start of the planning period covered by the power source plan to the end of the planning period is recorded and accumulated as actual information.

The learning task 160 uses, for example, machine learning using a neural network to create a neural network including an intermediate layer in which the profit rate factors are the input layer and the profit rate is the output layer, based on data obtained from performance information recorded and accumulated in the performance database 132.

Then, the learning task 160 creates an intermediate layer using actual profit rates as the results of power source maintenance carried out in the past as learning data (teaching data), and learns various parameters in the intermediate layer. The learning task 160 learns various parameters in the intermediate layer so as to minimize the error between the predicted profit rate and the actual profit rate included in the learning data (teaching data).

Learning task 160 predicts the profitability of each new power source based on some or all of the items output from tasks 151 to 157 above, or profitability factors for which other items are set, in accordance with the prediction model constructed as described above.

The learning task 160 further determines whether to install or cancel each new power source based on the profitability predicted as described above.

Specifically, for each new power source, the learning task 160 determines that the project should be newly installed if the profitability predicted in the manner described above is equal to or greater than a predetermined profitability threshold, and determines that the project should be canceled if the profitability is less than the predetermined profitability threshold.

The above-mentioned predetermined profitability threshold is set to the same value as the profitability threshold used in the judgment process by judgment task 159.

Then, the learning task 160 outputs the rate of return (ERm in FIG. 2) for each new power source, as well as the decision on whether to install or cancel (Jm in FIG. 2).

As shown in the example in FIG. 2, the rate of return calculated by the expected rate of return calculation task 158 (ER in the figure) may differ from the rate of return predicted by the learning task 160 (ERm in the figure). As the rate of return values differ, the judgment results (J and Jm in the figure) may also differ.

For power plant α in the example shown in Figure 2, the rate of return (ER) calculated by expected rate of return calculation task 158 is 9.2, and the rate of return (ERm) predicted by learning task 160 is 8.9. If the rate of return threshold is set to 8, then both the judgment result (J) by judgment task 159 based on the calculation result by expected rate of return calculation task 158 and the judgment result (Jm) based on the prediction result by learning task 160 will be "new construction."

For power plant β in the example shown in Figure 2, the rate of return (ER) calculated by expected rate of return calculation task 158 is 8.4, and the rate of return (ERm) predicted by learning task 160 is 7.8. If the rate of return threshold is set to 8, the judgment result (J) by judgment task 159 based on the calculation result by expected rate of return calculation task 158 will be "new construction," while the judgment result (Jm) based on the prediction result by learning task 160 will be "cancelled."

In this way, by providing a judgment result (J) by the judgment task 159 based on the calculation result by the expected return calculation task 158 and a judgment result (Jm) based on the prediction result by the learning task 160, the two results can be interpreted as complementary and used as useful information in formulating power supply plans.

As described above, the power supply plan development support device 1 and the power supply plan development support program 19 determine how to handle a power supply based on the profit calculated by subtracting the total of the fuel cost and fixed cost of the power supply from the total of the power volume revenue and the capacity revenue. This makes it possible to use the profit as an evaluation index to evaluate the profitability of each power supply and provide information on the change in profit when each existing power supply is maintained or abolished, or when a new power supply is newly established.

The power supply plan development support device 1 and power supply plan development support program 19 also determine how to handle a power supply based on the rate of return calculated using the revenue obtained by subtracting the total of the fuel cost and fixed cost of the power supply from the total of the power volume revenue and capacity revenue, and the initial investment amount of the power supply. This makes it possible to evaluate the profitability of each power supply using the rate of return as an evaluation index and provide information on the change in revenue when each existing power supply is maintained or abolished, or when a new power supply is newly established.

In addition, when the power supply plan formulation support device 1 and the power supply plan formulation support program 19 estimate the market price of electricity using a prediction model constructed by machine learning with the power price factors as input and the power price as output, it becomes possible to estimate a more realistic price of electricity that reflects actual information on past electricity prices, which in turn makes it possible to improve the reliability of the technology for supporting the formulation of power supply plans.

In addition, when the power supply plan formulation support device 1 and the power supply plan formulation support program 19 estimate the profitability of a power source using a prediction model constructed by machine learning with profitability factors as input and profitability as output, it becomes possible to estimate a more realistic profitability that reflects actual information on past profitability, which in turn makes it possible to improve the reliability of the technology for supporting the formulation of power supply plans.

The above describes an embodiment of the present invention, but the specific configuration is not limited to the above embodiment, and even if there are design changes that do not deviate from the gist of the present invention, they are included in the present invention.

For example, in the above embodiment, both judgments are made about existing power sources and judgments about new power sources, but it is also possible to make only one of judgments about existing power sources and judgments about new power sources. Note that when only judgments are made about existing power sources, the configuration does not include expected return calculation task 158.

In addition, in the above embodiment, in addition to the judgment task 159, the learning task 160 is included, but it is also possible to include only one of the judgment task 159 and the learning task 160. By providing the judgment results of the judgment task 159 for a new power source and the judgment results of the learning task 160 for a new power source in a mutually complementary manner, the accuracy of the assumptions and estimates related to the formulation of the power source plan is improved, and the profitability of the power source can be evaluated with a higher accuracy, thereby further improving the reliability of the technology for supporting the formulation of the power source plan. However, it is also possible to support the formulation of the power source plan by providing only the judgment results of either the judgment task 159 or the learning task 160.

REFERENCE SIGNS LIST 1 Power supply plan formulation support device 11 Input unit 12 Display unit 13 Storage unit 131 Power supply database 132 Performance database 14 Memory 15 Main task 151 Energy amount estimation task 152 Price estimation task 153 Energy amount revenue calculation task 154 Capacity revenue estimation task 155 Fuel cost estimation task 156 Fixed cost estimation task 157 Revenue calculation task 158 Expected rate of return calculation task 159 Judgment task 160 Learning task 17 Communication unit 18 Central processing unit 19 Power supply plan formulation support program

Claims (8)

A means for calculating the power revenue of a power source by multiplying the amount of power generated by the power source by a market price of the power source;
A means for calculating a capacity revenue of the power source by multiplying the power generation surplus of the power source by a supply capacity price in the market;
means for calculating revenue of the power source by subtracting a sum of fuel costs and fixed costs of the power source from a sum of the energy revenue and the capacity revenue;
and a means for determining that the existing power source should be continued if the revenue is equal to or greater than a predetermined revenue threshold , and for determining that the existing power source should be discontinued if the revenue is less than the predetermined revenue threshold .
A power supply plan formulation support device characterized by: A means for calculating the power revenue of a power source by multiplying the amount of power generated by the power source by a market price of the power source;
A means for calculating a capacity revenue of the power source by multiplying the power generation surplus of the power source by a supply capacity price in the market;
means for calculating revenue of the power source by subtracting a sum of fuel costs and fixed costs of the power source from a sum of the energy revenue and the capacity revenue;
means for calculating a rate of return for a new power source using the return and an initial investment amount of the power source;
means for determining that the new power source should be installed when the rate of return is equal to or greater than a predetermined rate of return threshold , and for determining that the new power source should be discontinued when the rate of return is less than the predetermined rate of return threshold ;
A power supply plan formulation support device characterized by: a means for estimating the electric energy price in the market using a prediction model constructed by machine learning with an electric energy price factor that fluctuates the electric energy price as an input and an electric energy price as an output;
3. The power supply plan formulation assistance device according to claim 1 or 2. a means for estimating the rate of return of the new power source using a prediction model constructed by machine learning with a rate of return factor having a significant effect on the rate of return as an input and a rate of return as an output;
3. The power supply plan formulation assistance device according to claim 2. Computer,
A means for calculating the power revenue of a power source by multiplying the amount of power generated by the power source by the power price in the market;
A means for calculating a capacity revenue of the power source by multiplying the power generation surplus of the power source by the supply capacity price in the market;
means for calculating revenue of the power source by subtracting a sum of fuel costs and fixed costs of the power source from a sum of the energy revenue and the capacity revenue; and
When the revenue is equal to or greater than a predetermined revenue threshold, the system determines that the existing power source should be continued, and when the revenue is less than the predetermined revenue threshold, the system determines that the existing power source should be discontinued.
A power supply planning support program characterized by: Computer,
A means for calculating the power revenue of a power source by multiplying the amount of power generated by the power source by the power price in the market;
A means for calculating a capacity revenue of the power source by multiplying the power generation surplus of the power source by the supply capacity price in the market;
means for calculating revenue of the power source by subtracting a sum of fuel costs and fixed costs of the power source from a sum of the energy revenue and the capacity revenue;
means for calculating a rate of return for a new power source using the revenue and an initial investment amount for the power source; and
when the rate of return is equal to or greater than a predetermined rate of return threshold, it is determined that the new power source should be newly installed, and when the rate of return is less than the predetermined rate of return threshold, it is determined that the new power source should be discontinued.
A power supply planning support program characterized by: Computer,
and functioning as a means for estimating the electricity price in the market using a prediction model constructed by machine learning with an electricity price factor that fluctuates the electricity price as an input and an electricity price as an output.
7. The power supply plan formulation support program according to claim 5 or 6. Computer,
a prediction model constructed by machine learning using a profit rate factor having a significant effect on the profit rate as an input and a profit rate as an output, the prediction model functioning as a means for estimating the profit rate of the new power source;
7. The power supply plan formulation support program according to claim 6.