JP6520228B2 - Power demand forecasting system, power demand forecasting method and program - Google Patents

Description

  The present invention relates to a power demand forecasting system, a power demand forecasting method, and a program.

  In recent years, in the power market, it has become possible to freely trade power (buy and sell power). In the case of conducting a power transaction, the trading price of the power is determined for a certain future period, and the power is traded (for example, trades every 30 minutes two days ago).

  From such a background, there has been proposed a method and the like for predicting the amount of power demand on a trading day (for example, two days later) so that the power supplier can manage risk in the power transaction (see Patent Document 1)

JP, 2004-252967, A

  From the background as described above, data relating to the amount of power demand in the past has been accumulated, and it has become possible to predict in advance the amount of power demand in a plant facility etc. accurately to some extent. However, the predicted value of the power demand calculated as described above has a fluctuation range. In the power transaction etc., in order to calculate the risk, it is desirable to accurately grasp the fluctuation range of the predicted value of the power demand amount.

The main invention of the present invention for solving the above-mentioned problems is a system for predicting the demand amount of electric power, which stores a first storage unit for storing weather information, data representing the relationship between air temperature and the demand amount. A fourth storage unit for storing data representing a relationship between sunshine duration and the demand amount, a fifth storage unit for storing past data on measured air temperature , and a power demand amount A sixth storage unit for storing past data, a seventh storage unit for storing past data on sunshine duration, an air temperature prediction unit for calculating probability distribution data of predicted air temperature based on the weather information; A sunshine time prediction unit for acquiring a predicted value of sunshine time, data representing a relationship between the air temperature and the demand amount, data representing a relationship between the sunshine time and the demand amount, probability distribution data of the air temperature forecast, and the sunshine Based on the predicted values between, and a demand prediction unit for calculating a probability distribution data of the demand, the demand prediction unit, and historical data on the measured temperature, the historical data relating to the demand, Based on the past data on the sunshine time, the temperature and the difference between the predicted air temperature and about 18 degrees Celsius, the square of the difference between the predicted air temperature and about 18 degrees Celsius, and the sunshine time as explanatory variables; calculating a regression equation between sunshine duration and demand, based on the predicted value of the regression equation and the probability distribution data of the predicted temperature the daylight hours, it characterized that you calculate the probability distribution data of the demand I assume.

  Further, in the power demand amount prediction system of the present invention, the sunshine duration prediction unit may acquire probability distribution data of a predicted value of the sunshine duration as a predicted value of the sunshine duration.

Further, the power demand forecasting system of the present invention is a system for forecasting a demand for power, and a performance value storage unit for storing performance values of air temperature for each time zone, and normalizing the performance values to calculate the time Reference curve creation unit that creates a reference curve that represents the change in temperature of each band, a first storage unit that stores weather information, and a third storage unit that stores data that represents the relationship between the temperature and the demand amount A fourth storage unit for storing data representing the relationship between sunshine hours and the demand amount, and calculating probability distribution data of predicted temperatures based on the weather information, and based on the maximum temperature and the minimum temperature and the reference curve Temperature forecasting unit for computing predicted temperatures for each time zone, a maximum / minimum temperature obtaining unit for obtaining past forecast values and future forecast values of the maximum temperature and the minimum temperature, and A prediction error calculation unit for giving the prediction temperature of the past to the temperature prediction unit by giving the prediction value of the past of the air temperature and the minimum temperature, and calculating a prediction error of the prediction temperature of the past; and prediction of the sunshine time A sunshine time prediction unit for acquiring a value, data representing the relationship between the air temperature and the demand amount, data representing the relationship between the sunshine time and the demand amount, probability distribution data of the predicted air temperature, and a predicted value of the sunshine time A demand forecasting unit that computes probability distribution data of the demand amount, and the air temperature forecasting unit calculates the forecasted air temperature in the future based on the highest temperature, the lowest temperature, and the reference curve, The probability distribution data of the predicted temperature may be calculated based on the predicted temperature of the future and the prediction error .

Another aspect of the present invention is a method of predicting a demand for power, which comprises the steps of: storing weather information; storing data representing a relationship between air temperature and the demand; Storing data representing a relationship between time and the demand amount, storing past data regarding measured air temperature, storing past data regarding power demand, and storing past data regarding sunshine duration Calculating the probability distribution data of predicted air temperature based on the weather information, acquiring the predicted value of the sunshine time, and data representing the relationship between the air temperature and the demand amount, the sunshine time and the demand The probability of the demand amount based on the data representing the quantity relationship, the probability distribution data of the predicted temperature, and the predicted value of the sunshine time Calculating the cloth data, and calculating the probability distribution data of the demand amount, the past step regarding the measured air temperature, the past data regarding the demand amount, and the past regarding the sunshine time Based on the data, at least the difference between the predicted air temperature and about 18 degrees Celsius, the square of the difference between the predicted air temperature and about 18 degrees Celsius, and the sunshine hours as the explanatory variables The probability distribution data of the demand amount is calculated based on the regression expression, the probability distribution data of the predicted air temperature, and the predicted value of the sunshine time.

Another aspect of the present invention is a program for predicting a demand for power, which comprises storing weather information in a computer, and storing data representing the relationship between air temperature and the demand. Storing the data representing the relationship between the sunshine hours and the demand amount, storing the past data on the measured air temperature, storing the past data on the power demand, and the past data on the sunshine hours Storing the temperature of the weather, calculating the probability distribution data of the predicted temperature based on the weather information, acquiring the predicted value of the sunshine duration, and data representing the relationship between the temperature and the demand, the sunshine duration And data representing a relationship between demand amount, probability distribution data of the predicted temperature, and the predicted value of the sunshine time, A program for executing the step of calculating probability distribution data of the required amount, wherein the step of calculating the probability distribution data of the demand amount includes: past data regarding the measured air temperature; and a past regarding the demand amount Of the difference between the predicted temperature and approximately 18 degrees Celsius, the square of the difference between the predicted temperature and approximately 18 degrees Celsius, and the sunshine time As a, the regression equation of the temperature and the sunshine time and the demand amount is calculated, and the probability distribution data of the demand amount is calculated based on the regression equation, the probability distribution data of the predicted temperature, and the predicted value of the sunshine time. .

Further, another aspect of the present invention is a method of predicting a demand amount of power, wherein a computer stores a temperature value of an air temperature for each time zone, and the performance value is normalized to obtain the time value. Creating a reference curve representing a change in air temperature, storing weather information, storing data representing a relationship between air temperature and the demand amount, data representing a relationship between sunshine duration and the demand amount Storing the probability distribution data of predicted air temperature based on the weather information, calculating the predicted air temperature for each time zone based on the highest temperature and the lowest temperature, and the reference curve; And obtaining respective past predicted values and future predicted values of the lowest temperature, and the past prediction of the highest temperature and the lowest temperature. A step value and based on the reference curve to calculate the past the prediction temperature of each of the time zones, and calculates a prediction error of the prediction temperatures of the past, obtaining a predicted value of the daylight hours, the air temperature And the data representing the relationship between the demand amount, the data representing the relationship between the sunshine duration and the demand amount, the probability distribution data of the predicted air temperature, and the predicted value of the sunshine duration, the probability distribution data of the demand amount is calculated Performing the step of calculating the predicted air temperature for each of the time zones, calculating the predicted air temperature in the future based on the maximum air temperature, the minimum air temperature, and the reference curve; Probability distribution data of the predicted air temperature is calculated based on the prediction error.

Further, another aspect of the present invention is a program for predicting a demand amount of electric power, which comprises the steps of storing in a computer a performance value of temperature in each time zone, and normalizing the performance value to calculate the time A step of creating a reference curve representing a change in air temperature for each zone, a step of storing weather information, a step of storing data representing a relationship between air temperature and the demand amount, a relationship between sunshine duration and the demand amount Storing data, calculating probability distribution data of predicted air temperature based on the weather information, and calculating predicted air temperature for each of the time zones based on the maximum temperature and the minimum temperature and the reference curve; Obtaining a past forecast value and a future forecast value of each of the maximum temperature and the minimum temperature; Serial calculates past the predicted temperature of each past predicted value and the time zone based on the reference curve, a step of calculating a prediction error of the prediction temperatures of the past, obtaining a predicted value of the daylight hours The data representing the relationship between the temperature and the demand, the data representing the relationship between the sunshine time and the demand, the probability distribution data of the predicted temperature, and the probability distribution of the demand based on the predicted value of the sunshine time Calculating a data, wherein the step of calculating the predicted air temperature for each of the time zones is the future predicted air temperature based on the maximum air temperature, the minimum air temperature, and the reference curve. Is calculated, and probability distribution data of the predicted air temperature is calculated based on the future predicted air temperature and the prediction error.

  Other problems disclosed in the present application and their solutions will be made clear by the section of the embodiments of the present invention and the drawings.

  According to the present invention, the fluctuation range of the predicted value of the power demand can be accurately grasped.

It is a figure showing a prediction system in an embodiment of the present invention. It is a figure which shows the hardware constitutions of the prediction apparatus in 1st Embodiment of this invention. It is a figure which shows the hardware constitutions of the transaction information provision apparatus in 1st Embodiment of this invention. It is a figure which shows the data table regarding the electric power generation means in 1st Embodiment of this invention. It is a figure which shows the data table regarding the working priority of the electric power generation means in 1st Embodiment of this invention. It is a figure which shows the data table regarding the transaction price in 1st Embodiment of this invention. It is a figure which shows the function structure of the prediction apparatus in 1st Embodiment of this invention. It is a figure which shows the flow of operation | movement of the prediction system in 1st Embodiment of this invention. It is an image figure of probability distribution in a 1st embodiment of the present invention. It is a figure which shows the marginal cost line in 1st Embodiment of this invention. It is a figure which shows the marginal cost line in 1st Embodiment of this invention. It is a figure which shows the marginal cost line in 1st Embodiment of this invention. It is a figure which shows the flow of operation | movement of the air temperature estimation part in 1st Embodiment of this invention. It is a figure which shows the data table regarding the past measurement air temperature in 1st Embodiment of this invention. It is a figure which shows the data table regarding the past predicted air temperature in 1st Embodiment of this invention. It is a figure which shows the data table regarding the future predicted air temperature in 1st Embodiment of this invention. It is an image figure of probability distribution of prediction air temperature in a 1st embodiment of the present invention. It is a figure which shows an example of probability distribution data of the prediction air temperature in 1st Embodiment of this invention. It is a figure which shows the result of the regression analysis in 1st Embodiment of this invention. It is a figure which shows the flow of operation | movement of the demand forecasting part in 1st Embodiment of this invention. It is a figure which shows the data table regarding the past electric power demand amount in 1st Embodiment of this invention. It is a figure which shows the flow of operation | movement of the demand forecasting part in 2nd Embodiment of this invention. It is a figure which shows the data table regarding the electric power demand amount in the past in 2nd Embodiment of this invention. It is a figure which shows the data table regarding the prediction reference value of the future amount of electric power demand in 2nd Embodiment of this invention. It is a figure explaining variable transformation of probability distribution in other embodiments of the present invention. It is a figure which shows an example of the data table M10 regarding the measured value of sunshine hours. It is a figure which shows an example of the data table M11 regarding the predicted value of sunshine hours. It is a figure which shows an example of the predicted value of the electric power demand which the demand forecasting part 103 calculates. The structural example of the prediction system which concerns on 5th Embodiment is shown. It is a figure which shows the software structural example of the air temperature prediction apparatus 10 which concerns on 5th Embodiment. 5 is a diagram showing an example of the configuration of a performance value storage unit 131. FIG. It is a figure which shows the structural example of the prediction value memory | storage part 132. FIG. It is a figure which shows the structural example of the reference | standard curve memory | storage part 133. FIG. FIG. 6 is a diagram showing an example of the configuration of a weight storage unit 134. FIG. 7 is a diagram showing an example of a configuration of an error data storage unit 135. It is a figure explaining the predicted value of the time slot | zone of a unit of 30 minutes. It is a figure which shows the flow of the pre-processing performed before performing air temperature prediction. It is a figure which shows the flow of preparation processing of a reference | standard curve. It is a figure which shows the flow of past temperature prediction processing. It is a figure which shows the flow of an analysis process of the prediction error by the error analysis part 115. FIG. It is a figure which shows the flow of future temperature prediction processing. It is a figure which shows the flow of the process which produces probability distribution of the air temperature predicted value for every time slot. It is a figure which shows the flow of the process which calculates the predicted value of the time slot | zone of a unit of 30 minutes. It is a figure which shows an example of the predicted value of the air temperature output by the predicted value output part 117. FIG.

  At least the following matters will be made clear by the present specification and the description of the accompanying drawings.

First Embodiment
=== About the prediction system ===
In the present embodiment, the probability of the temperature of the predetermined time zone is based on the understanding that the amount of the actual power demand fluctuates from the predicted value of the power demand based on the fluctuation of the air conditioning demand due to the temperature fluctuation. From the distribution, the probability distribution of the power demand amount in the relevant time zone is calculated.

  And the prediction system which this embodiment mentions later calculates the probability distribution of the profit value at the time of performing a fixed quantity of power transactions based on the probability distribution of the amount of power demand of the time slot concerned.

  FIG. 1 shows an example of a prediction system for realizing the prediction of the profit amount in the power transaction in the present embodiment. The prediction system according to the present embodiment includes the prediction device 100, the transaction information providing device 200, and the like. These devices transmit and receive data using the communication network 300 by LAN connection or the like.

  The prediction device 100 is a computer operated by the user. In addition, the transaction information providing device 200 is a computer that transmits data in response to a request from the prediction device 100. The prediction device 100 according to the present embodiment is configured to predict a profit amount by acquiring transaction information related to a transaction price and the like from the transaction information providing device 200.

  FIG. 2A shows the hardware configuration of the prediction device 100 of the present embodiment.

  The prediction device 100 includes a control unit 100A, a storage unit 100B, a communication unit 100C, an input unit 100D, and a display unit 100E.

  The control unit 100A is a CPU or the like, and is connected to the storage unit 100B, the communication unit 100C, the input unit 100D, and the display unit 100E via a bus or the like. Then, based on the computer program stored in the storage unit 100B, the control unit 100A performs data communication with the storage unit 100B, the communication unit 100C, the input unit 100D, and the display unit 100E, and controls their operation.

  The storage unit 100B includes volatile memory (RAM), non-volatile memory (flash memory), and the like. The storage unit 100B stores a computer program for controlling the prediction device 100, data to be described later that associates the amount of power demand and the power generation cost, and the like. The storage unit 100B also includes a storage unit (not shown) that stores a regression model described later, intermediate data calculated by each functional unit, final data, acquired analysis target data, and the like.

  The communication unit 100C is a communication controller or the like, and transmits / receives data to / from the transaction information providing apparatus 200 using the communication network 300 or the like by LAN connection by wire or wireless.

  The input unit 100D is a switch, a touch panel, or the like, and receives a user's operation instruction to the prediction device 100.

  The display unit 100E is a liquid crystal display or the like that displays various types of information, and displays a probability distribution or the like of a predicted profit amount described later.

  FIG. 2B shows a hardware configuration of the transaction information providing device 200 of the present embodiment.

  The transaction information providing apparatus 200 includes a control unit 200A, a storage unit 200B, and a communication unit 200C.

  The control unit 200A is a CPU or the like, and is connected to the storage unit 200B and the communication unit 200C via a bus or the like. Then, based on the computer program stored in the storage unit 200B, the control unit 200A performs data communication with the communication unit 200C and the storage unit 200B, and controls those operations.

  The storage unit 200B includes volatile memory (RAM), non-volatile memory (flash memory), and the like. The storage unit 200B stores a computer program for controlling the transaction information providing apparatus 200, data M3 related to transaction information to be described later, and the like.

  The communication unit 200C is a communication controller or the like, and transmits / receives data to / from the prediction device 100 using a communication network 300 or the like by LAN connection by wire or wireless.

  The data for correlating the power demand amount and the power generation cost stored in the storage unit 100B of the prediction device 100 of the present embodiment is calculated based on a data table M1 for power generation means described later and a data table M2 for operation priority of the power generation means Be done.

  FIG. 3A shows an example of the data table M1 regarding the power generation means of the present embodiment.

  The data M1 on the power generation means are the minimum power generation amount (MWh / h), the power generation amount (MWh / h), and the marginal cost (yen / MWh) for the plurality of power generation means owned by the power supplier.

  A power supplier has a plurality of power generation means such as hydroelectric power generation, solar power generation, nuclear power generation, and thermal power generation. The marginal cost, the amount of power that can be generated, and the like differ depending on the type of power generation means and the performance such as the power generation efficiency. For example, since nuclear power generation is difficult to stop operation, it will be scheduled as a means of power generation to always operate. On the other hand, in thermal power generation, the marginal cost may rise depending on the fuel price and the like. Besides, hydroelectric power generation is also characterized in that its marginal cost is cheaper than other power generation means.

  The power generation means A, B, C,... Shown in the data table of FIG. 3A represent one of those power generation means. In addition, the minimum amount of power generation (for example, 1 MWh / h) is difficult to stop operation from the viewpoint of the power generation cost, etc., as in the mass transit power source (the above-mentioned nuclear power generation) Represents the amount of power generated by Also, the power generation capacity (for example, 1 MWh / h) is a limit value of the power generation capacity of each of the power generation means. The marginal cost (yen / MWh) represents the power generation cost required per unit power generation (for example, 1 MWh).

  Here, the marginal cost may be set differently depending on the operation rate, that is, the power generation capacity being exhibited among the limit values of the power generation capacity of the power generation means, and the average of the operation rate 0 to 100% You may set a value. In the data table of FIG. 3A, the marginal costs of the power generation means A, B, C, ... are described as A (x 1), B (x 2), C (x 3), ... This represents a marginal cost function according to the operating rates x1, x2, x3. The storage format may be a table format or the like.

  FIG. 3B shows an example of the data table M2 related to the operation priority of the power generation means in the present embodiment.

  The data M2 relating to the operation priority of the power generation means has two types of priority relating to the mass transit power source and priority relating to the marginal cost. The priority for the mass transit power source is set higher than the priority for the marginal cost. In addition, the priority for marginal costs is set to operate in order from the one with the lowest marginal cost. In the above embodiment, the priority relating to the mass transit power source and the priority relating to the marginal cost are described as data relating to the operating priority, but it may be an aspect in which only the priority relating to the marginal cost is set. Priority may be set. In addition, data M2 related to the operation priority of each power generation means may be set by the user in consideration of the situation of each power generation means, etc., and when calculating the amount of profit, the profit prediction unit 104 or the like may be set.

  FIG. 3C shows an example of the data table M3 related to the transaction information stored in the storage unit 200B of the transaction information providing device 200 of the present embodiment.

  Data M3 on transaction information is data on the transaction price per unit power generation amount (for example, 1 MWh / h) in a predetermined time zone (for example, 11 o'clock to 12 o'clock) which is the object of the transaction, and data on the tradable amount . Here, the data related to the transaction information may be a fixed value or a predicted value. The data table M3 stores the transaction price and the tradable amount per hour.

  Power trading is conducted based on the transaction price. Specifically, when a power supplier sells a predetermined amount of power, based on the transaction price, the power supplier earns the sale instead of supplying the sale extra power. Can be obtained as Similarly, when the power supplier buys a predetermined amount of power, the power supplier uses the power supply for the buy instead of receiving the power supply for the buy from the transaction target person based on the transaction price. It is an aspect of paying the amount of money. Also, the tradable amount represents the upper limit value that the power supplier can trade in the power transaction.

  The data relating to the tradable amount may be set by the operation convenience of the power generation means of the power supplier. For example, as described above, since it is premised that the mass transit power supply is always operated, the power due to the operation of the mass transit power supply may be set as untradable. In addition, data on tradable volume need not necessarily be set.

  FIG. 4 shows an example of a functional configuration of the prediction device 100 of the present embodiment.

  The prediction device 100 includes an acquisition unit 101, an air temperature prediction unit 102, a demand prediction unit 103, and a profit prediction unit 104 described below according to the computer program stored in the storage unit 100B and the above hardware configuration (100A to 100E). , Implement the function of the presentation unit 105.

  The acquisition unit 101 communicates with other devices such as the transaction information providing device 200 to acquire data.

  The air temperature prediction unit 102 calculates data on the probability distribution of predicted air temperature in a predetermined time zone in the future, based on the predicted highest temperature, predicted lowest temperature, weather information and the like predicted by the Japan Meteorological Agency.

  The demand prediction unit 103 calculates data on the probability distribution of the power demand in a predetermined time zone in the future based on the data on the probability distribution of the predicted air temperature and the relational expression of the temperature and the power demand.

  The profit prediction unit 104 calculates, as a probability distribution, the change cost of the power generation cost when performing a power transaction of a predetermined amount based on the data on the probability distribution of the power demand amount and the data for correlating the power demand amount and the power generation cost. Do. And the transaction price at the time of performing a predetermined amount of power transaction is calculated, and the difference with the change expense of the power generation cost concerned is computed as probability distribution of the amount of profits of power transaction. Here, the data for correlating the power demand amount and the power generation cost is, for example, data on a marginal cost line calculated using data M1 on the power generation means and data M2 on the operation priority of the power generation means.

  The presentation unit 105 performs predetermined image processing on the probability distribution data of the profit amount, and presents the probability distribution of the predicted profit in such a manner that the user of the prediction device 100 can recognize the probability distribution. As an example, the presentation unit 105 displays, on the display unit 100E, the range between the minimum amount and the maximum amount of profit that can be taken as actual values with a probability of 95% as the probability distribution of the amount of profit.

  The data on probability distribution in the present embodiment (hereinafter referred to as “probability distribution data”) indicates the variation of the value that can be realized from the predicted value, and, for example, a predetermined future time zone (for example, 2013) It means the predicted temperature, the amount of power demand, and the range of the amount of profit (confidence interval) that can be realized with a probability of 95% at 7 o'clock on 2/1/1). Further, the probability distribution data includes data representing a probability density function related to a predicted value, data representing a probability distribution function, data related to a confidence interval calculated from the cumulative probability of samples, correspondence between the probability that can be realized and the width of the predicted value Data indicating a variance, a variance coefficient related to a predicted value, or the like.

=== About the operation of the prediction system ===
Next, the operation of the prediction system will be described.

  FIG. 5 shows an example of a flowchart of the present embodiment.

  (S1) is a step in which the user of the prediction device 100 inputs a period to be predicted (for example, 7 o'clock to 8 o'clock of 2013/2/1).

  (S2) is a process in which the prediction device 100 acquires transaction information from the transaction information providing device 200. Specifically, the acquisition unit 101 of the prediction device 100 requests the transaction information providing device 200 about the acquired price and the tradeable amount for the period to be predicted. Then, in response to the request, the transaction information provision device 200 transmits the transaction price and the tradeable amount regarding the predetermined time slot to be predicted from the data table M3 regarding the transaction information to the prediction device 100.

  In (S3), the temperature prediction unit 102 of the prediction apparatus 100 calculates the predicted value of the temperature of the predetermined time zone to be predicted as the probability distribution, based on the predicted maximum temperature predicted by the Japan Meteorological Agency, predicted minimum temperature, weather information, etc. Process. Here, when the target period of the power transaction is a long time spanning several hours, the predicted value of the air temperature for a plurality of corresponding time zones may be calculated. FIG. 6 shows an image of the probability distribution of predicted air temperature. FIG. 6 shows the probability distribution of predicted temperatures when one explanatory variable X is used for easy understanding.

  In addition, this embodiment is based on the understanding that the dispersion | variation in prediction air temperature leads to the dispersion | distribution of electric power demand amount, and the calculation method of probability distribution of prediction air temperature is arbitrary if a fixed precision. An example of a method of calculating the probability distribution of predicted air temperatures will be described later.

  (S4) is a predetermined time zone to be predicted based on the probability distribution data of predicted air temperature calculated by (S3) by the demand prediction unit 103 of the prediction device 100, the day of the week information, the power demand on the same month last year, etc. Calculating the predicted value of the power demand amount as a probability distribution.

  In the present embodiment, based on the understanding that the amount of actual power demand fluctuates from the predicted value of the amount of power demand based on the understanding that the fluctuation of the air conditioning demand caused by the fluctuation of the air temperature From the probability distribution, the probability distribution of the power demand in the relevant time zone is calculated. As a result, the probability distribution of the power demand can be accurately calculated. The method of calculating the probability distribution of the power demand amount is arbitrary as long as it has a certain accuracy, and data representing the relationship between the air temperature and the power demand amount may be stored in advance as a data table. An example of the method of calculating the probability distribution of demand forecast will be described later.

  In (S5), the profit prediction unit 104 determines a plurality of power generation means in a period subject to the power transaction based on the data M1 on the power generation means, the priority on the mass transit power source, and the data M2 on the priority It is a process of operating as it is, that is, calculating data on marginal cost lines. The data on marginal cost lines is an example of data for correlating the power demand amount with the power generation cost, and how to use the power generation means to compensate the power generation amount in correspondence with the required power generation amount (power demand amount) It is the data specified about In addition, when the data regarding a marginal expense line are stored beforehand, the process of (S5) can be skipped.

  The data on the marginal cost line calculated in (S5) will be described below in the form of a graphed marginal cost line.

  FIG. 7A is a marginal cost line calculated from data M1 on the power generation means and data M2 on the operation priority of the power generation means. The horizontal axis represents the required amount of power generation (the amount of power demand), and the vertical axis represents the marginal cost of each power generation means. The rightward direction indicates that the amount of power generation (the amount of power demand) is large, and the upward direction indicates that the marginal cost is large. Also, the section under the marginal cost line represents the difference in the means of power generation to be operated (the same applies to FIGS. 7B and 7C). Here, since the priority relating to the mass transit power source is set as a higher priority than the priority relating to the marginal cost, the marginal cost line is composed of two regions, the mass transit power source priority area A and the marginal cost priority area B There is. In other words, this means that the power generation means A, B, C are operated earlier than the power generation means D with lower marginal cost.

  FIG. 7B shows the relationship between the predicted value of the amount of power demand and the power generation means to be operated. FIG. 7B shows that the power generation unit F is operated when the amount of power demand is N.

  (S6) is a step in which the profit prediction unit 104 calculates the change cost of the power generation cost when the power transaction of the predetermined amount C is performed.

  The profit prediction unit 104 generates power when the power transaction of the predetermined amount C is performed based on the probability distribution data of the power demand amount calculated in the process of (S4) and the data regarding the marginal cost line calculated in (S5) Calculate the cost of change.

  Here, with reference to FIG. 7B and FIG. 7C, the change cost of the power generation cost when the power transaction of the predetermined amount C is performed will be described. FIG. 7B shows the power generation cost required when the amount of power demand is N. The total value of the areas of the shaded areas in FIG. 7B is the required power generation cost. Therefore, the power generation cost can be expressed by equation (1) at this time.

(F (W) represents a marginal cost line indicating the marginal cost corresponding to the power generation W)
Here, F (W) is calculated based on data M1 on the marginal cost of each power generation means, operation priority, and the like.

  Moreover, FIG. 7C shows the change cost (reduction) of the power generation cost when the transaction of the transaction amount C (power purchase) is performed in the period targeted for the power transaction. As shown in FIG. 7C, the change cost of the power generation cost at the time of buying the power of the trading amount C is a region where F (W) is surrounded by the power demand amount N of the W axis and the post-trade N-C. . Specifically, the change cost of the power generation cost can be expressed by equation (2).

In the present embodiment, since the power demand amount N is calculated as a probability distribution, the change cost of the power generation cost is also represented as a probability distribution.

  (S7) is a process in which the profit prediction unit 104 calculates the balance of transaction when the power transaction of a predetermined amount is performed in the period targeted for the power transaction based on the data M3 related to the transaction information. Specifically, when the transaction price in the period subject to the power transaction is a constant value R, the transaction balance (payment amount) in the case of performing a transaction (a power purchase) of a predetermined amount C is Can be represented.

(S8) is the change cost of the power generation cost when the profit prediction unit 104 trades the predetermined amount calculated in (S6), and the transaction balance when the predetermined amount of trade calculated in (S7) is performed Is the step of calculating the probability distribution of the amount of profit by the difference of Specifically, in the case of power purchase, the larger the difference between the change cost (reduction) of the generation cost and the balance of the transaction (payment amount), the larger the profit for the power supplier. That is, the profit amount Y of the power supplier in the case of the power purchase can be expressed as Formula (4) from Formula (2) and Formula (3).

When the tradable amount is set, the transaction amount C is set with the tradable amount as the upper limit value.

  Here, as described above, the power demand amount is data including probability distribution. From this, the probability distribution of the profit amount Y can be calculated based on the probability distribution data of the power demand amount. The steps (S6) and (S7) may be omitted because they are substantially integrated into the step (S8).

  The specific contents of the method of calculating the probability distribution of profit will be described later.

  (S9) is a process in which the presentation unit 105 performs predetermined image processing and presents the probability distribution of the profit amount in the power transaction calculated in (S8) so that the user of the prediction device 100 can recognize it. For example, the presentation unit 105 displays, as text data, the width or the like of the predicted profit that can be realized with a probability of 95% as the probability distribution of the predicted profit when performing a power transaction of a predetermined amount.

  As described above, according to the present embodiment, the user of the present system can grasp the probability distribution of the profit amount in the power transaction in consideration of the fluctuation of the power demand amount caused by the temperature fluctuation.

  Further, in the above embodiment, the case of power purchase has been described, but the same applies to the case of power sale. That is, in the case of the power seller, the difference between the balance of transactions (the amount of income) and the cost of change (increase) of the generation cost is the benefit for the power supplier. At this time, the profit amount Y can be expressed as Expression (5).

Further, in the above embodiment, the marginal cost line is calculated based on the data M1 on the power generation means and the data M2 on the operation priority of the power generation means as data for correlating the power demand amount and the power generation cost, We calculated the change cost of the power generation cost of the case. However, the data for correlating the power demand amount with the power generation cost may be data etc. stored in a table format, correlating the power demand amount with the power generation cost.

  Further, in the above embodiment, the case where the transaction price R is constant regardless of the transaction amount C has been described. However, the transaction price R may be variable according to the transaction amount C. In that case, R in equation (4) can be expressed as a function R (C) of C.

  Further, in the above embodiment, although the prediction system calculates the amount of profit, it is not limited to the amount of profit but may be predetermined points or the like as long as it is data relating to the value of profit.

=== About the operation of the temperature prediction unit ===
Next, an example of the operation of the air temperature prediction unit 102 will be described.

  In the present embodiment, the temperature is predicted based on the understanding that the temperature is related to the predicted maximum temperature of the day, the predicted minimum temperature, and the time zone at that time. That is, from the measurement result of the past temperature, the predicted maximum temperature and the predicted minimum temperature, regression analysis is performed to calculate the relationship between the temperature in the time zone and the predicted maximum temperature and the predicted minimum temperature as a regression equation. Then, the probability distribution of predicted air temperature is calculated from the standard deviation s of the error term in the regression analysis.

  FIG. 8 shows a flowchart of this embodiment.

  In the present embodiment, the prediction device 100 uses the weather information prediction device 400 and the communication network 300 such as LAN connection (not shown in FIG. 1) to measure actual temperatures and predicted maximum temperatures regarding past days. By performing transmission and reception of data such as predicted minimum temperature, predicted maximum temperature and predicted minimum temperature regarding future days, temperature prediction is performed.

  The weather information prediction apparatus 400 is a computer that transmits data in response to a request from the prediction apparatus 100. Further, the weather information prediction apparatus 400 has a hardware configuration similar to that of the transaction information providing apparatus 200 shown in FIG. 2B.

  FIGS. 9A to 9C show an example of the air temperature related data table stored in the storage unit of the weather information prediction apparatus 400. FIG. 9A is the actual temperature of the past day, FIG. 9B is the predicted maximum temperature of the past day, the predicted minimum temperature, and FIG. 9C is the data about the predicted maximum temperature of the future, the predicted minimum temperature.

  FIG. 9A is a data table M4 of the air temperature measured every hour. In the data table M4, the measurement result of the air temperature measured at a predetermined time (hereinafter, referred to as "measured value") is stored in association with the measurement date and time. Here, the measurement result is, for example, one obtained by measuring the temperature of the outside air with a thermometer on an hourly basis. The measurement date is the date and time when the measurement was made.

FIG. 9B is a data table M2 of the temperature predicted by the Japan Meteorological Agency or the like. In this data table M5, the predicted maximum temperature T1 _max and the predicted minimum temperature T1 _min predicted by the Japan Meteorological Agency etc., and the data regarding the past days are stored in correspondence with the days on which the target of the forecast was made ing. The measurement date of the air temperature of the data table M4 may be stored in a format corresponding to the date of the data table M5, and may be stored in a format associated with the same data table. Also, instead of the date, it may be stored in a form associated with a predetermined code. Further, the time when the measurement is performed may also be stored in the form of a predetermined timing or the like associated with the time of day.

FIG. 9C is a data table M6 of the temperature predicted in advance by the Japan Meteorological Agency etc. for the day to be predicted. In this data table M6, predicted maximum temperature T2 _max and predicted minimum temperature T2 _min predicted by the Japan Meteorological Agency etc., and data on future days are stored in association with the target dates.

In (S31) of FIG. 8, the acquisition unit 101 of the prediction device 100 causes the weather information providing device 400 to measure, among the past data, the measured temperature measured during a predetermined time zone that is the target of the power transaction, This is a process for requesting the predicted maximum temperature T1 _max and the predicted minimum temperature T1 _min on the measurement date.

In (S32), the weather information providing device 400 receives the request, acquires the measured value measured in the corresponding predetermined time period from the data table M4 related to the measured value of the past data, , And transmitting the data. In addition, the weather information provision device 400 predicts the predicted maximum temperature T1 _max corresponding to the measurement date of the actual measurement value from the data table M5 related to the predicted maximum temperature and predicted minimum temperature predicted by the Japan Meteorological Agency etc. in advance of past data. The minimum temperature T1 _min is acquired, and the data is transmitted to the prediction device 100.

As an example, when the target period of the power transaction is 7 o'clock to 8 o'clock in 2013/2/1, the acquiring unit 101 determines that the actual value measured at 7 o'clock or 8 o'clock in the data M4 related to the measured temperature in the past Obtain the predicted maximum temperature T1 _max and the predicted minimum temperature T1 _min of the measurement date.

  In the present embodiment, the data on the time zone for which a prediction is desired, which is acquired from the past data in (S31), is a range in which there is no problem in clarifying the relationship between the time zone, the predicted maximum temperature, and the predicted minimum temperature. It means inside. Therefore, for example, when the time slot to be predicted is 7 o'clock to 8 o'clock, the past data showing 7 o'clock even if the past measured value is data measured at 7:15 or 6:45. It becomes an acquisition target. Similarly, the meaning of “time zone” in the present invention does not necessarily indicate time, such as 14:00 to 15:00, and a predetermined section in the transition of the time of day. means.

(S33) is a process in which the air temperature prediction unit 102 of the prediction device 100 calculates the air temperature prediction formula based on the acquired past data. Specifically, the prediction device 100 executes regression analysis based on the actual measurement value T1, the predicted maximum temperature T1 _max , and the predicted minimum temperature T1 _min , and calculates the temperature prediction equation relating to the predicted temperature T2 in the predetermined time zone. At the same time, the probability distribution of the predicted temperature T2 is calculated.

  The regression analysis is performed by, for example, the least squares method with respect to the regression model of Equation (6) using the predicted maximum temperature and the predicted minimum temperature as explanatory variables.

(Β ₀ represents the maternal intercept, β ₁ , β ₂ the maternal regression coefficient, E _i represents the error term, and i at the end of each variable represents each observation point i, and the past data obtained as a sample Indicates each actual value, predicted maximum temperature, predicted minimum temperature.)
From this, β ₀ , β ₁ , and β ₂ of the regression model are determined, and regression equation (7) is calculated as a temperature prediction equation.

(^ At the top of T is the notation of T in the regression equation)
Further, at this time, in order to calculate the probability distribution of the predicted air temperature, the standard deviation s of the error term is calculated based on the equation (8) as a coefficient relating to the variation of the predicted air temperature.

(Ti ^ represents the value of T when the predicted maximum temperature T1 _max and the predicted minimum temperature T1 _min are substituted into regression equation (7) for each observation point i.)
Then, in the present embodiment, assuming that the distribution of each actually measured value follows the normal distribution, the probability distribution of the predicted air temperature T2 uniformly varies the standard deviation s ² of the error term regardless of the value of the explanatory variable. It can be regarded as a normal distribution.

In this case, the probability distribution of the predicted air temperature T2 can be expressed from the standard deviation s of the error term of the regression equation (7) and the equation (8). For example, the probability distribution of the predicted air temperature T2 can be expressed as a probability density function f (T3) as shown in equation (10).

  FIG. 10 shows an image of the probability distribution of the predicted temperature T2.

(T3 represents the temperature deviated from the calculated predicted temperature T2)
Further, the probability distribution of the predicted air temperature T2 can also be calculated as a confidence interval. For example, the confidence interval can be calculated based on the interval integral of the probability density function f (T3). In this case, the 68.2% confidence interval is the temperature between the predicted air temperature T2 and the air temperature T2 ± s that is shifted by s.

  The shaded area in FIG. 10 represents the relationship between the 68.3% confidence interval and the probability density function f (T3).

(S34) is a process in which the acquisition unit 101 of the prediction device 100 requests the weather information providing device 400 to generate the predicted maximum temperature T2 _Max and the predicted minimum temperature T2 _min for the target date of 2013/2/1. is there.

In (S35), the weather information providing apparatus 400 receives the request, and the predicted maximum temperature T2 _max and the predicted minimum temperature T2 _min predicted by the Japan Meteorological Agency etc. are predicted target dates from data M6 regarding future days This is a step of acquiring the predicted maximum temperature T2 _max and the predicted minimum temperature T2 _min regarding (for example, 2013/2/1), and transmitting the data to the prediction device 100.

  (S36) is based on the prediction data acquired by the air temperature prediction unit 102 of the prediction device 100, the calculated air temperature prediction formula, and a coefficient relating to the variation of the predicted air temperature T2, a predetermined time slot of the prediction target date (for example, It is a process of calculating predicted temperature T2 of 7 o'clock to 8 o'clock of 2013/2/1 and its probability distribution data.

  As described above, the air temperature prediction unit 102 according to the present embodiment can calculate the probability distribution of predicted air temperature from the weather information with high accuracy.

  The probability distribution data may be calculated as a confidence interval corresponding to the probability, in which case the standard deviation s can be calculated by multiplying a coefficient corresponding to the reliability. For example, in the case of the reliability 70%, the standard deviation s may be multiplied by a coefficient 1.28, and in the case of the reliability 90%, the standard deviation s may be multiplied by a coefficient 1.53.

In addition, as the probability distribution data, a confidence interval may be calculated by using a cumulative distribution function in past data. FIG. 11 is an example in which the cumulative probability is calculated by using a plurality of samples acquired from the past data, and the confidence interval is calculated. Column of the error represents the value of the error term E _i for regression equation (7), the column of frequency sample number corresponding to the error, the column of the probability density probability density calculated from the number of samples and frequency, the cumulative probability The columns represent the cumulative probability calculated from the number of samples and the frequency. For example, FIG. 11 shows that the range of 0.8 ° C. error from the value predicted from the regression equation is a 70% confidence interval.

  Moreover, in the said embodiment, when the time slot | zone of a power transaction is 1 hour of 7 o'clock-8 o'clock of 2013/2/1, it describes that the temperature prediction part 102 calculates the regression about 7 o'clock-8 o'clock did. However, when dealing with long hours, such as when the time zone of the power transaction is 3 hours from 7 o'clock to 10 o'clock of 2013/2/1, 7 o'clock to 8 o'clock, 8 o'clock to 9 o'clock, 9 o'clock to 9 o'clock Regression equations for 10 o'clock etc may be calculated separately.

  Moreover, in the said embodiment, the regression model which calculates the prediction air temperature which makes only the prediction highest temperature and the prediction lowest temperature the explanatory variable was used. However, if the probability distribution of predicted air temperature can be calculated with a certain degree of accuracy from weather information, the regression model does not have to be limited to the above. For example, a regression model in which only one of the predicted maximum temperature and the predicted minimum temperature is used as an explanatory variable may be used. As weather information to be referred to in order to calculate the probability distribution of predicted temperatures, predicted maximum temperatures, predicted minimum temperatures, weather information, wind speed information, atmospheric pressure and the like can be mentioned. On the other hand, seasonal information, area information, etc. may be added as explanatory variables in order to improve the prediction accuracy. Also, in order to stabilize the dispersion of the sample, a dispersion stabilization transformation may be performed to adapt the regression model.

  An example of a regression equation reflecting seasonal information is shown in equation (11).

(Β ₀ is an intercept, β ₁ is a regression coefficient, ν _j is a seasonal adjustment coefficient for each season)
According to the regression equation (11), it is possible to calculate the predicted temperature with the influence of the predicted maximum temperature T _max and the predicted minimum temperature T _min appropriately adjusted with the influence by the season. Note that FIG. 12 shows an example of the result of the confidence interval of the predicted air temperature calculated for each time by regression (11).
Further, since the temperature prediction unit 102 aims to calculate the probability distribution of the power demand amount by accurately grasping the probability distribution of the predicted temperature, the calculation method of the probability distribution of the other known predicted temperatures is It may be adopted.

=== About the operation of the demand forecasting unit ===
Next, an example of the operation of the demand prediction unit 103 will be described.

  In the present embodiment, the prediction of the amount of power demand is based on the understanding that the fluctuation of the heating and cooling demand due to the temperature change is one of the factors that the actual amount of power demand changes from the predicted value of the amount of power demand. Do. That is, from the probability distribution of the predicted air temperature calculated by the air temperature prediction unit 102, the probability distribution of the power demand amount is calculated. Then, in the present embodiment, the demand prediction unit 103 calculates the relationship between the temperature and the amount of power demand based on the actual value of the temperature and the actual value of the amount of power demand in the past, thereby obtaining the probability distribution of the amount of power demand. Calculate

  FIG. 13 shows an example of a flowchart of the present embodiment.

  In this embodiment, the prediction device 100 uses the weather information providing device 400, the demand information providing device 500, and the communication network 300 such as a LAN connection (not shown in FIG. 1) to measure the actual temperature, and By performing transmission and reception of past data and the like regarding the power demand amount at the time, the power demand amount is predicted.

  The demand information providing device 500 is a computer that transmits data in response to a request from the prediction device 100. Further, the demand information providing device 500 has the same hardware configuration as the transaction information providing device 200 shown in FIG. 2B.

  FIG. 14 shows an example of a data table M7 related to the amount of power demand actually measured and stored in the storage means of the demand information providing device 500. In the data table M7, the power demand in a predetermined time zone is associated with the date and time, and is stored on an hourly basis.

  (S41) in FIG. 13 is a process in which the acquiring unit 101 of the prediction device 100 requests the demand information providing device 500 for the past amount of power demand related to the predetermined time slot that is the target of the power transaction.

  (S42) is a step in which the demand information providing device 500 receives the request, acquires the power demand amount from the data table M7 related to the past power demand amount, and transmits the data to the prediction device 100. It is.

  As an example, when the target period of the power transaction is 7 o'clock to 8 o'clock in 2013/2/1, the acquisition unit 101 is 7 o'clock to 8 o'clock in the data M7 related to the amount of power demand in the past (the same month last year). Get data The time zone and the like of the power demand amount to be acquired can be appropriately changed in consideration of the similarity to the target period of the power transaction and the like.

  (S43) is a process in which the acquiring unit 101 of the prediction device 100 requests the weather information providing device 400 to measure the past air temperature regarding the time zone of the power demand acquired in (S42).

  (S44) is a step in which the weather information providing device 400 receives the request, acquires an actual measurement value from the data table M4 related to the past measured air temperature, and transmits the data to the prediction device 100. .

  As an example, when the time zone of the amount of power demand acquired in (S42) is 7 o'clock to 8 o'clock of 2012/2/1 to 2012/2/8, the acquiring unit 101 is 2012/2/1 to 2012 The data of 7 o'clock or 8 o'clock of data M4 related to the measured temperature of 2/8 is acquired.

  (S45) is a process in which the demand prediction unit 103 of the prediction device 100 calculates a demand prediction formula based on the acquired past data. Specifically, the prediction apparatus 100 is a regression analysis for calculating the relational expression between the amount of power demand and the temperature based on the actual value T1 of the temperature in the predetermined time zone and the amount of power demand N1 acquired in (S44). I do.

  The regression analysis is, for example, the regression model of the equation (12) in which the power demand N is the objective variable, the difference between the standard temperature 18 ° C. and the actual measurement value, and the square of the difference between the standard temperature 18 ° C. and the actual measurement value , By the least squares method.

(Γ ₀ , δ ₀ are population intercepts, γ ₁ , γ ₂ , δ ₁ , δ ₂ are population regression coefficients, E _i is an error term, and i at the end of each variable represents each observation point i , Represents each actual measurement value T1 of the past data acquired as a sample, and the amount of power demand N1.)
Here, the difference between the standard temperature of 18 ° C and the actual measurement value (air temperature) is taken as an explanatory variable because the amount of fluctuation of the power demand due to the heating and cooling demand is substantially zero when the standard temperature is 18 ° C. It can be regarded as. Also, by taking the square of the difference between the standard temperature 18 ° C and the actual measurement (air temperature) as an explanatory variable, the demand for air conditioning and heating will increase from the general social phenomenon that the temperature rapidly increases as it deviates from the standard temperature 18 ° C. It can be accurately reflected.

  Further, the equation (12) divides the equation into two depending on whether it is 18 ° C. or more or 18 ° C. or less, and the cooling demand and the heating demand are separated. Although omitted in the equation (12), regression analysis may be performed by adding explanatory variables such as weather information and day information.

From this, γ ₀ , γ ₁ , γ ₂ , δ ₀ , δ ₁ and δ ₂ of the regression model are determined, and regression equation (13) is calculated as a demand forecast equation.

(S46) is a step in which the demand prediction unit 103 of the prediction device 100 calculates the probability distribution of the power demand based on the above regression equation (13) and the probability distribution data of the predicted temperature calculated by the temperature prediction unit 102. It is.

  As an example, when the probability distribution data of the predicted temperature is represented by the expected value T2 of the predicted temperature and the temperature T2 ± s as a confidence interval of 68.2%, the probability distribution (reliable interval) of the power demand amount is It becomes as follows.

That is, the expected value of the amount of power demand is calculated by substituting the expected value T2 of the probability distribution of the predicted air temperature into the air temperature T of Expression (13). Then, the upper limit value and the lower limit value of the confidence interval of the predicted value of the power demand amount can be calculated by substituting T2 + s and T2-s into equation (13) (equation (13A) is for temperature T) It can be regarded as a monotonically decreasing function.Equation (13B) can be regarded as a monotonically increasing function with respect to the temperature T.)
The expected value T0 of the probability distribution of predicted air temperatures can be calculated from equation (14) when the probability density function of the predicted air temperatures is f (T).

As described above, the demand prediction unit 103 can calculate the probability distribution of the power demand amount with high accuracy.

  In the present embodiment, a regression model is used in which the difference between the standard temperature 18 ° C. and the actual measurement (air temperature) and the square of the difference between the standard air temperature 18 ° C. and the actual measurement (air temperature) are explanatory variables. However, if the probability distribution of the power demand can be calculated with a certain degree of accuracy from the probability distribution of the predicted temperature, the regression model does not have to be limited to the above. For example, for the square of the difference between the standard temperature 18 ° C and the actual measurement (air temperature), the explanatory variable may be omitted, or instead of the difference between the standard temperature 18 ° C and the actual measurement (air temperature) Temperature) may be used as an explanatory variable. The standard temperature may be approximately 18 ° C., and may be set to 17 ° C. or 19 ° C. Also, predicted maximum temperature, predicted minimum temperature, weather information, regional information, day of the week information, etc. may be added as explanatory variables.

  Further, in the above embodiment, when calculating the demand forecasting equation, the demand information providing device 500 is configured to acquire past data at the same time as the forecast target date. However, any other method may be used as long as it is a method of calculating the air temperature prediction formula based on past data at the same time as the time of the prediction target date. For example, a variable relating to time is added as a dummy variable to the regression model of equation (12) in advance, and other time information is excluded when the demand prediction unit 103 of the prediction device 100 performs regression analysis. It is also good. In this case, it may be set to be 1 at the same time and 0 otherwise.

  Also, in order to stabilize the dispersion of the sample, a dispersion stabilization transformation may be performed to adapt the regression model. In the regression model of Formula (12) which calculates a demand forecast formula, an example at the time of performing dispersion | distribution stabilization conversion is shown to Formula (15).

(Γ ₀ represents a maternal intercept, γ ₁ ... Γ _{N + 4} represents a maternal regression coefficient, E _i represents an error term, and D _n represents a dummy variable corresponding to a day of the week, a day or a holiday)
In the case of the regression model of equation (12), the variation tends to increase depending on the temperature value, but according to the regression model of equation (15), the variation can be made uniform.

  Further, the purpose of the demand forecasting unit 103 is to calculate the probability distribution of the profit amount by accurately grasping the probability distribution of the power demand amount. Therefore, instead of the regression equation, the probability of the known power demand amount is calculated. A distribution calculation method may be adopted.

=== About the operation of the profit forecasting unit ===
Next, the operation of the profit prediction unit 104 will be described.

  In the present embodiment, the profit prediction unit 104 calculates probability distribution data of the profit amount Y based on the probability distribution data of the amount of power demand calculated by the demand prediction unit 103.

  The method of calculating the profit amount Y is as described in the above equation (4) or (5). In the following, a method of calculating the probability distribution of the profit amount Y will be described in the case where the probability distribution data of the power demand amount is the expected value N0 of the power demand amount and the data N2 to N3 regarding the confidence interval.

  The expected value of the profit amount Y can be calculated by substituting the expected value N0 of the power demand amount for the power demand amount N of equation (4).

  Further, the upper limit value Y2 and the lower limit value Y3 of the confidence interval of the profit amount Y can be calculated by substituting N2 and N3 respectively into the equation (4) (equation (4) is the power when C is constant) It can be regarded as a monotonically increasing function of the demand amount N).

  As described above, the profit prediction unit 104 can calculate the probability distribution of the profit amount based on the probability distribution of the power demand amount.

  As described above, according to the present embodiment, it is possible to reflect the probability distribution of predicted air temperatures calculated in units of time on the probability distribution of profit. That is, it is possible to set the aspect in which the fluctuation of the amount of power demand due to the air conditioning demand on a time basis is reflected in the probability distribution of the profit amount.

  Thus, the system user can recognize the probability distribution of the amount of profit when performing a certain amount of power transaction.

  In the above embodiment, although the trading volume in the power transaction is set to the predetermined volume C, C ′ at which the profit Y becomes maximum may be calculated as the optimal trading volume based on the probability distribution of the profit Y. In that case, for example, C ′ at which Y becomes maximum may be calculated when the predicted value N of the power demand amount is the expected value N0.

Second Embodiment
In the present embodiment, a method of calculating the probability distribution for the predicted reference value of the power demand amount will be described when the predicted reference value of the power demand amount is presented by the central power supply command center or the like.

  The central power supply command center or the like may calculate a forecast reference value of the amount of power demand based on past results, weather information, etc., and present it to the power supplier. And the prediction reference value of the amount of power demand calculated in this way is information in which the high accuracy reflecting the past results is secured.

  However, even with the predicted value of the amount of power demand calculated in this manner, as described in the first embodiment, if there is a change in temperature, it fluctuates due to a change in air conditioning and heating demand. is there.

  So, in this embodiment, demand forecasting part 103 'computes the probability distribution, making use of the prediction standard value of the amount of electric power demand presented by the central electric power feeding order place etc. That is, demand forecasting part 103 'will clarify the relational expression of how much the amount of power demand fluctuates from a prediction standard value, when air temperature changes 1 degree from predicted air temperature based on past data . Then, the probability distribution of the power demand amount is calculated based on the relational expression and the probability distribution of the predicted temperature. In the present embodiment, a relational expression is calculated to calculate the fluctuation amount of the power demand amount when the air temperature deviates by 1 degree with respect to the expected value T0 of the probability distribution of the predicted air temperature.

  The configuration other than the demand prediction unit 103 'is the same as that of the first embodiment, and thus the description thereof is omitted.

  FIG. 15 shows an example of a flowchart showing the operation of the demand prediction unit 103 'of this embodiment.

  In this embodiment, the prediction device 100 uses the weather information providing device 400, the demand information providing device 500, and the communication network 300 such as a LAN connection (not shown in FIG. 1) to measure the actual temperature, and By performing transmission and reception of past data and the like regarding the power demand amount at the time, the power demand amount is predicted.

FIG. 16A shows an example of a data table M8 regarding the past amount of power demand stored in the storage means of the demand information providing device 500. In this data table M8, as the past data, the actual value of the power demand in the predetermined time zone, the power demand predicted for the predetermined time zone, and the air temperature (° C.) predicted for the predetermined time zone , Are associated with the date and time, and are stored on an hourly basis.
Further, FIG. 16B shows an example of a data table M9 related to the predicted value of the power demand (hereinafter referred to as “the predicted reference value of the power demand”) stored in the storage unit of the demand information providing device 500. In the data table M9, predicted values of the future power demand are stored in one-hour units in association with date and time.

  (S41 ') of FIG. 15 is a process in which the acquiring unit 101 of the prediction device 100 requests the demand information providing device 500 for the past amount of power demand related to the predetermined time slot that is the target of the power transaction. Here, the acquisition unit 101 requests the amount of power demand at the predicted temperature T2 among the past data.

  The reason why the regression equation is calculated for the predicted air temperature T2 is that the amount of change in the amount of power demand caused by the air conditioning and heating demand is different for each air temperature. For example, the amount of change in the power demand differs between when the predicted temperature 19 ° C. actually changes to 20 ° C. and when the predicted temperature 30 ° C. actually changes 31 ° C.

  In (S42 '), the demand information providing device 500 receives the request, and is predicted about the power demand of the predetermined time zone actually measured and the predetermined time zone from the data table M8 related to the power demand in the past. This is a step of acquiring the power demand amount and the air temperature predicted for the predetermined time period, and transmitting the data to the prediction device 100.

  As an example, when the target period of the power transaction is 7 o'clock to 8 o'clock in 2013/2/1, and the expected value of the predicted temperature is 20 degrees, the acquiring unit 101 obtains the power demand of the past (the same month of last year) Data on quantity M7 at 7 to 8 o'clock and the temperature when the predicted temperature is 20 degrees are acquired.

  (S43 ′) is a process in which the acquiring unit 101 of the prediction device 100 requests the weather information providing device 400 to measure the past air temperature regarding the time zone of the power demand acquired in (S42 ′). is there.

  (S44 ′) is a process in which the weather information providing device 400 receives the request, acquires an actual measurement value from the data table M4 related to the past measured air temperature, and transmits the data to the prediction device 100. is there.

  As an example, when the time zone of the amount of power demand acquired in (S 42 ′) is 7 o'clock to 8 o'clock of 2012/2/1 to 12/2/2/8, the acquiring unit 101 performs the process from 2012/2/1 to 2012/01 The 7 o'clock or 8 o'clock data of data M4 regarding the measured air temperature of 2012/2/8 is acquired.

  (S45'-1) is a process in which the demand prediction unit 103 'of the prediction device 100 calculates a demand prediction formula based on the acquired past data. Specifically, the prediction device 100 measures the measured power demand amount of the predetermined time zone acquired in (S42 ′) and (S44 ′), the power demand amount predicted for the predetermined time zone, and the predetermined time zone The amount of electricity demand changes from the predicted value of electricity demand when the air temperature changes from 1 oC to 1 oC based on the air temperature predicted for the above and the measured value of the air temperature for a predetermined time period Perform regression analysis.

  The regression analysis is a regression model of equation (16) using the actual value of the power demand Nf as the objective variable, the difference between the predicted air temperature T0 and the actual measurement Tf, and the square of the difference between the predicted air temperature T0 and the actual measurement Tf as the explanatory variable. For the least squares method. In addition, since it is a regression model for calculating the deviation value from the predicted value of the amount of power demand, it is an arithmetic expression for the predicted value N0 of the amount of power demand.

(Γ ₁ , γ ₂ , δ ₁ , δ ₂ are population regression coefficients, E _i is an error term, N 0 is a predicted value of power demand, N f is a measured value of power demand, T 0 is a predicted temperature, Tf represents the measured value of the air temperature, and i at the end of each variable represents each observation point i, and the measured value of the air temperature of the past data obtained as a sample, the measured value of the power demand, and the prediction of the air temperature Represents the forecasted value of electricity demand
In this embodiment, the difference between the predicted air temperature T0 and the actual measurement value Tf, and the difference between the predicted air temperature T0 and the actual measurement value Tf, in order to predict the fluctuation amount of the power demand due to the air conditioning demand when the air temperature changes 1 ° C from the prediction value. A quadratic regression model is used, with the square of x as the explanatory variable. Further, as in the case of the equation (12), the equation is divided into two depending on whether the temperature is 18 ° C. or more or 18 ° C. or less.

From this, γ ₁ , γ ₂ , δ ₁ and δ ₂ of the regression model are determined, and regression equation (17) is calculated as a demand forecast equation.

(S45′-2) is a process in which the acquiring unit 101 of the prediction device 100 requests the demand information providing device 500 to estimate the power demand for the predetermined time slot that is the target of the power transaction.

  In (S45'-3), the demand information providing device 500 receives the request and predicts the power demand for the predetermined time period to be forecast from the data table M9 according to the forecast reference value of the future power demand. This is a step of acquiring a reference value and transmitting the data to the prediction device 100.

  In (S46), the demand prediction unit 103 ′ of the prediction device 100 calculates the probability distribution of the power demand based on the regression equation (17) and the probability distribution data of the predicted temperature calculated by the temperature prediction unit 102. It is a process.

  As an example, when the probability distribution data of the predicted air temperature is the expected value T0 and the air temperature T0 ± s shifted by s from the expected value T0, the probability distribution (reliable interval) of the power demand amount is calculated by equation (17) it can. That is, the upper limit value and the lower limit value of the confidence interval of the power demand amount are shifted from the expected value T0 of the predicted air temperature with respect to Tf while substituting the prediction reference value for the predicted value N0 of the power demand amount of equation (17) It can be calculated by substituting the temperature T0 ± s (Eq. (17) can be regarded as a monotonically increasing function of the temperature T).

  As described above, the demand prediction unit 103 'of the present embodiment can calculate the probability distribution with respect to the prediction reference value of the power demand amount with high accuracy.

  In the above embodiment, in the process of (S46) using the regression equation (17), the expected value T0 and the air temperature T0 ± s shifted by s from the expected value T0 are used as the probability distribution data of the predicted air temperature. The probability distribution of demand was calculated. However, instead of this, when the central power supply command center presents the predicted air temperature T0, the predicted air temperature T0 may be a value acquired from the demand information providing device 500 together with the prediction reference value of the amount of power demand. Further, as a deviation from the predicted air temperature T0, data on the confidence interval calculated from the cumulative probability shown in FIG. 11 may be used. Then, for each reliability (for each probability) of the predicted air temperature, a fluctuation range relative to the reference value of the amount of power demand may be calculated.

Incidentally, instead of the regression equation (17), using the following regression equation (18), the difference between the standard temperature 18 ° C. and the measured _{value, γ 1, γ 2, δ} 1, to calculate the [delta] ₂ It is also good.

In this case, the predicted value of the power demand for the past day can be obtained by setting Nf as the actual value of the power demand, Tf as the actual value of the past air temperature, and N0 as a virtual predicted value of the power demand. It is possible to eliminate the need for data on temperature and predicted temperature.

  Then, for the probability distribution (blur width) of the forecast reference value of the power demand amount, the forecast reference value of the power demand amount is substituted for N0 of the calculated regression equation (18) It can be calculated by substituting the temperature T0 ± s shifted from the value T0.

  In addition, when the predicted air temperature T0 is acquired from the demand information providing device 500 together with the predicted reference value of the amount of power demand (predicted air temperature of the central power supply command center), the prediction standard of the power demand amount when shifted from the predicted air temperature T0 to the air temperature Tf. The following equation (19B) can also be used as the probability distribution (blur width) of values as another aspect in the present embodiment.

(Formula (19A) is a regression equation of Formula (15), Formula (19B) represents a formula using the regression coefficient of Formula (15), and γ _{N + 1} , γ _{N +2} , γ _{N + 3} and γ _{N + 4} are Represents the regression coefficient calculated by equation (15).)
Gamma _{N + 1} of the formula _{(19B), γ N + 2} , γ N + 3, γ N + 4 are regression coefficients gamma _{N + 1} of the formula _{(15), γ N + 2} , γ N + 3, since using a gamma _{N + 4,} past power demand amount of the measured values And, it is possible to calculate the prediction formula easily from the past actual temperature value.

  Then, the probability distribution (blur width) of the forecast reference value of the power demand amount substitutes the forecast reference value of the power demand amount into N0 of the calculated regression equation (19B), and the demand information provision device for T0 It can calculate by substituting the temperature T0 +/- s (or the data regarding the reliability section of the above-mentioned predicted air temperature) which shifted from predicted air temperature T0 to predicted air temperature acquired from 500 and Tf.

In equation (19B), ΔDT ₁ , ΔDT ₁ ² , ΔDT ₂ , and ΔDT ₂ ² are used as explanatory variables when Tf deviates from predicted air temperature T0 of the central power supply command station, and the degree of influence on the amount of power demand Is because it can be expressed by ΔDT ₁ , ΔDT ₁ ² , ΔDT ₂ , and ΔDT ₂ ² .

The coefficient relating the amount of change in power demand amount when shifted 1 ℃ from the predicted temperature T0 (temperature sensitivity) already calculated and are gamma _{N + 1} in equation _{(15), γ N + 2} , γ N + 3, and gamma _{N + 4} since can be regarded as substantially the same, in the formula (19B), γ _{N + 1} is calculated by the formula _{(15), γ N + 2} , γ N + 3, it is used gamma _{N + 4.}

<Other Embodiments>
In the above embodiment, when the probability distribution of predicted air temperature is converted into a probability distribution of profit, a predetermined confidence interval calculated from the probability distribution of predicted air temperature is transformed into a confidence interval of power demand, and the power demand is calculated. We used a method to convert from confidence intervals of to confidence intervals of forecasted profits.

  However, the calculation method for converting the probability distribution of the predicted temperature into a variable and reflecting the result on the probability distribution of the predicted profit is arbitrary.

  For example, the probability density function of predicted air temperature is subjected to variable conversion based on regression equation (13) etc., and expressed as a probability density function of power demand, and similarly, the probability density function of power demand is probability of forecast profit It may be converted to a density function.

  FIG. 17 is an image diagram of variable conversion of the probability density function. When the probability density function f (x) is converted to the probability density function g (y), the variable-converted g (y) can be expressed as Expression (21) by transforming Expression (20).

In this case, it is possible to predict the temperature of the probability density function f (x) is not limited to a normal distribution, T distribution, chi ² distribution, F distribution, etc., apply to any distribution function.

  Similarly, the coefficient (standard deviation s of the error term) related to the variation of the predicted air temperature is converted into a coefficient related to the variation of the power demand based on the regression equation (13), and the coefficient related to the variation of the power demand May be variable-transformed into a coefficient relating to the variation of the predicted profit on the basis of the above equation (4).

  Further, in the above embodiment, the fluctuation of the demand for electricity is calculated as the largest fluctuation of the demand for heating and cooling due to the fluctuation of the air temperature as a factor that the actual amount of demand for electric power fluctuates from the predicted value of the amount of electricity demand. The error terms such as Equation (12), Equation (15), and Equation (16) are not considered. However, the standard deviation of the error term is calculated in the same manner as the standard deviation s of the error term of equation (8) also for the error terms such as equations (12), (15), and (16), It is good also as an aspect considered in calculating probability distribution of. In that case, the standard deviation may be integrated according to the well-known error propagation law.

Further, in the above embodiment, the probability distribution of the predicted air temperature is a normal distribution uniformly taking the standard deviation s ² of the error term as the population variance regardless of the values of the explanatory variables (predicted maximum air temperature and predicted low air temperature). I regarded it. However, the confidence interval of the predicted temperature may be calculated as a confidence interval based on the t distribution in which the distribution function is made different depending on the values of the explanatory variables (predicted maximum temperature and predicted minimum temperature). In that case, the coefficient relating to the variation of the predicted air temperature T2 can be expressed by equation (22) as the standard deviation s' of the error term as an unbiased estimator divided by n-3 degrees of freedom.

At this time, the distribution of the regression equation (7) can be estimated to follow the following normal distribution, and the confidence interval of the predicted temperature T2 can be calculated from the t value.

(The average value of each observation point is the value on regression equation (7), and the variance coefficient is determined based on the standard deviation s' of the error term. D ² is the predicted maximum temperature T2 _Max and predicted Represents the square of the Mahalanobis distance determined by the lowest temperature T2 _Min )
The confidence interval of the predicted air temperature T2 in this case can be calculated by the equation (24) based on the predicted highest temperature T2 _max , the predicted lowest temperature T2 _min , and the standard deviation s' of the error term.

The above calculation formula is omitted from the detailed description because it is a calculation method which is known in statistical terms.

Thus, when the probability distribution of the predicted air temperature T2 is represented as a confidence interval based on the t distribution, when the number of observation points is small, and when the observation points are biased, the predicted highest temperature T2 _max of the prediction target day, the predicted lowest Even in the case where the temperature T2 _min indicates a temperature range different from normal, etc., it is possible to calculate the probability distribution of predicted temperatures with high accuracy.

In the above embodiment, as the data indicating the probability distribution of the predicted temperature T2, may be obtained by using other distribution function, it may be data based on the chi ² distribution and F distribution.

  In the above embodiment, the least squares method is shown as a method of regression analysis. However, instead of the least squares method, the moment method or the maximum length method may be used.

  In the above embodiment, the air temperature prediction equation and the demand prediction equation are calculated according to the date and time of the prediction target being set, but the air temperature prediction equation and the demand prediction equation are generated in advance It is good also as a mode stored in storage means 100B, and according to the date and time of prediction object being set, the temperature prediction formula of a corresponding date and time, and a demand forecasting formula being acquired from storage means 100B.

  Moreover, in the said embodiment, it was set as the structure which the prediction apparatus 100 has the air temperature prediction part 102, the demand prediction part 103, and the profit prediction part 104 as a function part. However, these functional units or parts thereof may be distributed to other devices. Similarly, the storage area of data stored in each storage means may be any place. For example, it may be integrated in the prediction device 100, or may be distributed and stored on a cloud system composed of a plurality of computers.

Third Embodiment
In the present embodiment, the sunshine duration is considered in the prediction of the power demand. In the equation (12) of the first embodiment, the power demand amount N is used as the objective variable, and the difference between the standard temperature 18 ° C. and the actual measurement value and the square of the difference between the standard air temperature 18 ° C. and the actual measurement value In the present embodiment, the sunshine time is added to this as an explanatory variable. Specifically, the demand prediction unit 103 performs regression analysis by the least square method on the regression model of the following equation (12 ′) in (S45) of FIG.

Here, μ represents a mother regression coefficient, and S represents a sunshine time. In the present embodiment, it is assumed that the sunshine duration S is a value of 0.1 hour unit. FIG. 18 is a diagram showing an example of the data table M10 related to the actual measurement value of the sunshine duration. The measured value of the sunshine duration may be a time measured by a measuring instrument such as a sunshine meter, or a value provided by the meteorological agency or a meteorological company may be acquired.

  The power demand amount N determined by the equation (12 ') is the total demand amount by the customer minus the power generation amount by solar power generation, that is, the power demand amount for a power supplier such as a power company. The reason why sunshine duration is used as an explanatory variable is that the amount of power demand to the power supplier fluctuates due to the increase and decrease of the amount of power supply by solar power generation. This makes it possible to accurately predict the heating and cooling demand for the power supplier.

The acquiring unit 101 of the prediction device 100 transmits a request to the weather information providing device 400, and the weather information providing device 400 acquires the measured value of the air temperature from the data table M4 in response to the request, and the data table The measured values of the sunshine duration are acquired from M10 and transmitted to the prediction device 100. The demand prediction unit 103 of the prediction device 100 adds the γ ₀ , the γ ₁ , the γ ₂ , the δ ₂ , the δ ₀ , the δ ₁ , and the δ ₂ of the regression model using the actual temperature and the actual sunshine time. Determine μ and calculate the following regression equation (13 ′) as a demand forecast equation.

In (S46) of FIG. 13, the demand prediction unit 103 is based on the regression equation (13 ′), the probability distribution data of the predicted temperature calculated by the temperature prediction unit 102, and the probability distribution data of the predicted value of sunshine time. The probability distribution of electricity demand will be calculated. In the present embodiment, it is assumed that the predicted value of sunshine duration is given by manual input from the operator. For example, the demand forecasting unit 103 sets the expected value N0 of sunshine duration to the 50th percentile value, the lower limit value N2 of the confidence interval to the 10th percentile value, and the upper limit N3 to the 90th percentile value, and the median value 30th percentile value and 70 It can accept input of percentile values. A sunshine time prediction unit may be provided to input a predicted value of sunshine time.

  FIG. 19 is a diagram showing an example of the data table M11 related to the predicted value of sunshine duration. In the example of FIG. 19, the data table M11 includes 10th, 30th, 50th, 70th, and 90th percentile values of the predicted value S of sunshine duration for each time zone. Further, in the example of FIG. 19, the length of the time zone is 30 minutes, and the value of the predicted value of the sunshine duration is set for each of the 48 hour zones of one day. As shown in the figure, the 50th percentile value is the expected value of sunshine duration (as expected), 10th and 30th percentile values are set as sunshine durations below the forecast, and 70th and 90th percentile values as sunshine durations above the forecast. It is set.

  In addition, when the meteorological agency or the meteorological company provides the forecast value of the sunshine time, the forecast value of the sunshine time may be obtained as the forecast value of the sunshine time, or the sunshine time may be acquired based on the past weather data. May be predicted.

  FIG. 20 is a diagram illustrating an example of the predicted value of the amount of power demand calculated by the demand prediction unit 103. In the example of FIG. 20, the 10th, 30th, 50th, 70th, and 90th percentile values related to the prediction of the power demand amount are calculated in accordance with the input of the sunshine hours.

  Further, equation (15 ′) indicating a regression model in the case of performing dispersion stabilization conversion in equation (12 ′) may be used.

... Equation (15 ')
As described above, by adding the sunshine time as an explanatory variable, it is possible to accurately obtain the predicted value of the power demand for the power provider such as the power company, considering the decrease in the apparent demand due to the solar power generation. Can.

Fourth Embodiment
In the present embodiment, when the predicted reference value of the power demand amount is presented by the central power supply command center or the like, the probability distribution is calculated for the predicted reference value of the power demand amount while considering the sunshine time. Specifically, the demand prediction unit 103 ′ performs regression analysis using the following equation (16 ′) instead of the equation (16) in (S45′-1) of FIG.

S0 represents the predicted value of sunshine time, and Sf represents the actual value of sunshine time.

A regression analysis is performed by the least squares method for the regression model of equation (16 '). By this, when the air temperature changes from the predicted temperature T0 presented by the central power supply instruction document etc., how much the power demand amount changes from the predicted value of the power demand amount (γ ₁ , γ ₂ , δ ₁ , δ ₂ : Temperature sensitivity) and how much the power demand changes from the predicted value when the sunshine time changes from 0.1 to 0.1 hours (μ; sunshine time sensitivity) It can be asked.

By regression analysis, γ ₁ , γ ₂ , δ ₁ , δ ₂ and μ of the regression model of the above equation (16 ′) are determined, and the following equation (17 ′) is calculated as a demand forecast equation.

... Equation (17 ')
The demand prediction unit 103 ′ of the prediction device 100 calculates the power based on the above regression equation (17 ′), the probability distribution data of the predicted temperature calculated by the temperature prediction unit 102, and the probability distribution data of the input sunshine time. The probability distribution of demand can be calculated.

In addition, it replaces with the said regression equation (17), and calculates (gamma) ₁ , (gamma) ₂ , (delta) ₁ , (delta) ₂ , (mu) by difference of standard temperature 18 degreeC and actual value using following regression equation (18) You may

In addition, when the predicted air temperature T0 is acquired from the demand information providing device 500 together with the predicted reference value of the amount of power demand (predicted air temperature of the central power supply command center), the prediction standard of the power demand amount when shifted from the predicted air temperature T0 to the air temperature Tf The following equation (19B ′) can also be used for the probability distribution (blur width) of the values.

Here, ΔS = Sf−S0, Sf is the actual value of the amount of sunshine, and S0 is the predicted value of the amount of sunshine.

  As described above, it is possible to predict the amount of power demand in consideration of the amount of solar radiation.

Fifth Embodiment
In the present embodiment, the air temperature prediction device 10 predicts the air temperature, and the air temperature prediction unit 102 of the prediction device 100 acquires the air temperature prediction value from the air temperature prediction device 10. FIG. 21 shows one configuration example of the prediction system according to the present embodiment. As shown in the figure, the prediction device 100, the transaction information providing device 200, and the air temperature prediction device 10 are connected to the communication network 300, and transmit and receive data via the communication network 300.

  The air temperature prediction apparatus 10 according to the present embodiment normalizes the past actual temperature values to obtain a curve (hereinafter referred to as a reference curve) representing a change in air temperature in one day, and predicts air temperature based on the reference curve. . The normalization in the present embodiment is performed by obtaining an average value of actual values for each time zone, setting the minimum temperature of one day to “0”, and setting the maximum temperature to “1”. Therefore, the temperature prediction value for each time zone can be calculated by multiplying the reference curve by a value obtained by subtracting the minimum temperature from the maximum temperature and adding the minimum temperature.

In addition, the air temperature prediction device 10 of the present embodiment obtains the average value μ and the standard deviation σ (or variance σ ² ) of prediction errors of air temperature prediction based on the reference curve, and performs temperature distribution prediction in the future prediction. Calculated as

  In this embodiment, it is assumed that the temperature is predicted to predict the power demand in the Chugoku region, and the power supply target is Hiroshima, Okayama, Shimane, Tottori, Yamaguchi. It is assumed that it is five prefectures (hereinafter referred to as the forecast target area). Therefore, in the present embodiment, first, the air temperature of the above-described forecast target area is predicted, and then a weighted average value weighted according to the ratio of the power demand amount is used as the air temperature forecast value of the entire Chinese region.

=== Configuration of Temperature Prediction Device 10 ===
The air temperature prediction apparatus 10 has a hardware configuration similar to that of the prediction apparatus 100 shown in FIG. 2A.

  FIG. 22 is a diagram showing an example of the software configuration of the air temperature prediction apparatus 10 according to the present embodiment. The air temperature prediction apparatus 10 includes an actual value acquisition unit 111, a predicted value acquisition unit 112, a reference curve generation unit 113, an air temperature prediction unit 114, an error analysis unit 115, a probability distribution generation unit 116, an estimated value output unit 117, and an actual value storage unit. 131 includes a predicted value storage unit 132, a reference curve storage unit 133, a weight storage unit 134, and an error data storage unit 135.

  The actual value acquisition unit 111, the predicted value acquisition unit 112, the reference curve creation unit 113, the air temperature prediction unit 114, the error analysis unit 115, the probability distribution creation unit 116, and the predicted value output unit 117 are controlled by the air temperature prediction apparatus 10. It is realized by the means 100A reading and executing the program stored in the storage means 100B, and is realized by the actual value storage unit 131, the predicted value storage unit 132, the reference curve storage unit 133, the weight storage unit 134 and the error data storage unit 135. Can be realized as part of the storage area provided by the storage means 100B provided in the air temperature prediction apparatus 10.

  The actual value acquisition unit 111 acquires the actual value of the temperature. In the present embodiment, the actual temperature value is obtained by measuring the temperature of the outside air with a thermometer every predetermined time (one hour in the present embodiment, but may be any time), that is, time It will be the actual value for each band. In the present embodiment, the time zone is in units of one hour, and is divided into 24 segments from 1 o'clock to 24 o'clock. The actual value acquisition unit 111 can receive a measurement result automatically transmitted from, for example, a measurement device of air temperature, or can access the measurement device to acquire a measurement result. The actual temperature value may be, for example, a value provided by the Japan Meteorological Agency or a meteorological company. In this case, the actual temperature value can be obtained by accessing the computer of the meteorological agency or the meteorological company. The actual value acquisition unit 111 acquires actual temperature values for a plurality of measurement locations. As a measurement place, an area name to which a weather forecast is provided, that is, one of “Hiroshima”, “Okayama”, “Matsue”, “Tottori”, and “Yamaguchi” is set. The actual value acquisition unit 111 registers the acquired actual value of the temperature in the actual value storage unit 131.

  The actual value storage unit 131 stores the actual value of the temperature. FIG. 23 is a view showing a configuration example of the actual value storage unit 131. As shown in FIG. The actual value storage unit 131 stores an actual value for each measurement location in association with the measurement date (day), the time zone, and the season information. Of course, the measurement location can be set arbitrarily, but it should be the same as (or able to coincide with) the point at which the predicted values of the maximum temperature and the minimum temperature are provided. The season information represents season segments divided arbitrarily, and in the present embodiment, it is determined according to a predetermined rule according to the date and time, and the season information is "winter", "early spring", "spring", "early summer", "Saiyu" "Summer" "Autumn" shall be divided into seven divisions. The seasonal information may be set, for example, every other year for "Suiyu". In the actual value storage unit 131, actual values for the past 10 years are stored in advance. Note that a measurement result may be acquired from the measurement apparatus, or an actual value registration unit may be provided to receive an input of the measurement result from the operator and register the result in the actual value storage unit 131. Also, less than or more than 10 years worth of actual value may be registered.

  The predicted value acquisition unit 112 acquires predicted values of the maximum temperature and the minimum temperature of the prediction target day. The predicted value acquisition unit 112 can, for example, access a computer of the Meteorological Agency or a meteorological company to receive data (meteorological data) including a predicted value announced by the Meteorological Agency or the meteorological company. The predicted value acquisition unit 112 may receive input of predicted values of the maximum temperature and the minimum temperature from the operator.

  In the present embodiment, it is premised that the temperature forecasting after two days is performed to forecast the power demand after two days, and the forecasted value acquiring unit 112 calculates the maximum temperature and the minimum temperature two days after the two days before the forecast target day. Assume that a predicted value is to be obtained. Further, the predicted value acquisition unit 112 acquires predicted values of the maximum temperature and the minimum temperature for each area to be predicted. That is, the predicted value acquisition unit 112 acquires predicted values of the maximum temperature and the minimum temperature as weather forecasts for “Hiroshima”, “Okayama”, “Matsue”, “Tottori” and “Yamaguchi”. In the present embodiment, it is assumed that weather data including weather forecasted by the Japan Meteorological Agency, the probability of precipitation, and predicted values of the maximum temperature and the minimum temperature are acquired.

  The predicted value acquisition unit 112 periodically (for example, every day at a predetermined time) acquires weather data including predicted values of the maximum temperature and the minimum temperature two days later from the Meteorological Agency or a meteorological company. The predicted value acquisition unit 112 registers the acquired predicted value in the predicted value storage unit 132.

  The predicted value storage unit 132 stores predicted values of the maximum temperature and the minimum temperature. FIG. 24 is a view showing a configuration example of the predicted value storage unit 132. As shown in FIG. The predicted value storage unit 132 stores predicted values of weather, rainfall probability, maximum temperature, and minimum temperature in association with the name (area name) of the prediction target area and the prediction target date. The area name is either "Hiroshima", "Okayama", "Matsue", "Tottori" or "Yamaguchi". In the present embodiment, it is assumed that only one set of the predicted value of the highest temperature and the predicted value of the lowest temperature corresponding to the prediction target day is stored in the predicted value storage unit 132. As described above, the predicted value two days after announced in the past is registered in the predicted value storage unit 132. Further, in the present embodiment, it is assumed that predicted values for the past three years are stored in the predicted value storage unit 132.

  The reference curve creation unit 113 creates a reference curve. Specifically, the reference curve creation unit 113 records, for each date of one year and each region to be predicted, the actual temperature value corresponding to the region of the same month on the same month stored in the actual value storage unit 131, Make the lowest temperature "0" and normalize the highest temperature "1". More specifically, the reference curve creation unit 113 uses the actual value storage unit 131 for the same month on the same day of the same month on each day of the one year and each area (the leap year in the past ten years for February 29 for February 29). The actual value of the temperature of the number) is read out, the average value of the actual value of the temperature is calculated for each time zone, the lowest value of the average (hereinafter referred to as the average lowest temperature) and the highest temperature of the average (hereinafter referred to as the average highest Temperature is selected, and the average minimum temperature is subtracted from the average value of each actual value, and then a value obtained by subtracting the average minimum temperature from the average maximum temperature is calculated as a normalized value. It is a standard curve that arranged the normalization value for every time zone in order of time zone. The reference curve creation unit 113 registers the created reference curve in the reference curve storage unit 133.

  The reference curve storage unit 133 stores a reference curve for each prediction target area. FIG. 25 is a diagram showing a configuration example of the reference curve storage unit 133. As shown in FIG. As shown in FIG. 25, the reference curve storage unit 133 associates the temperature of each time zone from “0” to “1” for each date of one year in association with the name (area name) of the prediction target area. A normalized value of the average temperature (hereinafter referred to as normalized temperature) expressed by a numerical value between (in the example of FIG. 25, a percentage from 0% to 100%) is stored.

  The weight storage unit 134 stores weights for each area. FIG. 26 is a view showing a configuration example of the weight storage unit 134. As shown in FIG. In the present embodiment, the weight uses the ratio of the power demand in each area. The user may change the weight arbitrarily, or may set the weight regardless of the amount of power demand.

  The temperature prediction unit 114 predicts the temperature using the reference curve. The temperature prediction unit 114 is given the prediction target area and the highest and lowest temperatures, and the temperature prediction unit 114 takes a value obtained by subtracting the given lowest temperature from the given highest temperature as the given prediction target area. The temperature prediction value of the prediction target area is calculated by multiplying the numerical value of the corresponding reference curve and adding the lowest temperature given thereto. Further, the air temperature prediction unit 114 uses the weights stored in the weight storage unit 134 to perform weighted averaging of the air temperature prediction values of the area to be predicted, and calculates the air temperature prediction value of the whole (in this embodiment, the Chugoku region). .

  The error analysis unit 115 analyzes a prediction error relating to the prediction of the temperature using the reference curve. The error analysis unit 115 reads predicted values of the maximum temperature and the minimum temperature for the past date from the predicted value storage unit 132, gives the predicted values of the read maximum temperature and the minimum temperature to the temperature prediction unit 114, and obtains the predicted temperature. Calculate it. The error analysis unit 115 compares the predicted air temperature value calculated by the air temperature prediction unit 114 with the corresponding date and time zone actual value for each season information corresponding to the actual temperature value stored in the actual value storage unit 131. The difference (prediction error) is calculated, and the average value μ and the standard deviation σ of the prediction error are calculated for each of the season information and the time zone. The error analysis unit 115 registers the seasonal information and the average value μ and the standard deviation σ for each time zone in the error data storage unit 135.

  The error data storage unit 135 stores data on prediction errors (hereinafter referred to as error data). FIG. 27 is a view showing a configuration example of the error data storage unit 135. As shown in FIG. The error data storage unit 135 stores the average value μ and the standard deviation σ for each time zone in association with the season information (winter, early spring, spring, early summer, soup, summer, autumn 7 sections).

The probability distribution creation unit 116 creates data (hereinafter referred to as probability distribution data) that represents the predicted value predicted by the air temperature prediction unit 114 with a probability distribution. In this embodiment, assuming that the predicted value of the air temperature is a normal distribution, the predicted value of the air temperature is added to the average value (expected value) of the prediction error, and then the standard distribution σ is changed to the normal distribution N (μ, σ ² Create data (eg, probability density function f (x) of normal distribution) representing. In the present embodiment, the probability distribution generator 116 generates data representing the normal distribution N (μ, σ ² ) of the prediction error from the average value μ of the prediction error and the standard deviation σ, and generates the 10th percentile value for the prediction error. A probability distribution of predicted temperature values is created by calculating the 30th percentile value, the 50th percentile value, the 70th percentile value, and the 90th percentile value, and adding the predicted values of the temperature of the corresponding time zone.

  The predicted value output unit 117 outputs a predicted temperature value. The predicted value output unit 117 outputs a predicted temperature value based on the probability distribution created by the probability distribution creation unit 116. In this embodiment, the predicted value output unit 117 outputs five values of the 10th percentile value, the 30th percentile value, the 50th percentile value, the 70th percentile value, and the 90th percentile value of the probability distribution as predicted temperatures for each time zone. . The predicted value output unit 117 can also output the average value (expected value) μ as a predicted value of the air temperature, or can output a graph of normal distribution for each time zone. In the present embodiment, the prediction value output unit 117 transmits the prediction value to the prediction device 100, and the air temperature prediction unit 102 of the prediction device 100 acquires the prediction value transmitted by the prediction value output unit 117.

The predicted value output unit 117 can also create and output a predicted value in a time zone of 30 minutes from the predicted value in a time zone of 1 hour unit. As shown in FIG. 28, it is assumed that the predicted value of each time zone is the predicted value of the intermediate time of each time zone, and this is divided by the length of the time zone to predict a time zone of any length. You can determine the value. For example, assuming that the predicted value of air temperature in a certain time zone X in one hour unit is T _X , the predicted temperature T _XS of air temperature at the start time Xs of time zone X and the air temperature forecast at the start time (X + 1) S of time zone X + 1 The value T _{(X + 1) S} is obtained by the following equation.

Then, assuming that the time zone of the first half 30 minutes constituting the time zone X is XA and the time zone of the second half 30 minutes is XB, the predicted air temperature values T _XA and T _XB of the time zones XA and XB are determined by the following equations.

=== Processing ===
Hereinafter, the process performed by the air temperature prediction apparatus 10 of this embodiment is demonstrated.

<Pre-processing>
FIG. 29 is a diagram showing a flow of pre-processing performed before temperature prediction.

  The actual value acquisition unit 111 acquires the actual value of the air temperature and registers it in the actual value storage unit 131 (S201), and the predicted value acquisition unit 112 acquires the weather data and registers it in the predicted value storage unit 132 (S202) . The reference curve creation unit 113 creates a reference curve from the actual temperature value (S203).

  FIG. 30 is a diagram showing the flow of the process of creating a reference curve. The reference curve creation unit 113 first reads out the actual value of the temperature of the same area on the same month and the same day from the actual value storage section 131 for each area and each day of the year (including February 29) (S221) The actual temperature values are averaged for each time zone to calculate the average temperature for each time zone (S222). The reference curve creation unit 113 selects an average maximum temperature and an average minimum temperature from among the average temperatures (S223). The reference curve creation unit 113 subtracts the average lowest temperature from each average temperature (S224), and divides the subtracted average temperature by the value obtained by subtracting the average lowest temperature from the average highest temperature to calculate a normalized value (S225). The reference curve creation unit 113 registers, in the reference curve storage unit 133, a reference curve in which normalized values are arranged in order of time zone (S226).

  Referring back to FIG. 29, the temperature prediction unit 114 performs past temperature prediction for each region and date (S204).

  FIG. 31 is a diagram showing a flow of past temperature prediction processing.

  The air temperature prediction unit 114 performs the following processing for all the prediction target days stored in the prediction value storage unit 132, that is, each of the prediction target days for the past three years.

  The temperature prediction unit 114 calculates predicted values of the maximum temperature and the minimum temperature corresponding to the target area and the target date for each area of “Hiroshima”, “Okayama”, “Matsue”, “Tottori” and “Yamaguchi” from the predicted value storage unit 132 Reading (S241), the reference curve corresponding to the area and date is read from the reference curve storage unit 133 (S242). The air temperature prediction unit 114 multiplies the normalized value of each time zone included in the read reference curve by a value obtained by subtracting the predicted value of the lowest air temperature read out from the read out predicted value of the highest air temperature, The prediction value of the air temperature for each time zone is calculated by adding the prediction value of the minimum air temperature read out (S243).

  Referring back to FIG. 29, the error analysis unit 115 analyzes the prediction error (S205).

  FIG. 32 is a diagram showing a flow of analysis processing of a prediction error by the error analysis unit 115.

  The error analysis unit 115 calculates prediction target dates, regions, and all the prediction target dates stored in the prediction value storage unit 132, that is, each prediction target date for the past three years, each region, and each time zone. The actual temperature value and the seasonal information corresponding to the time zone are read out from the actual value storage unit 131 (S261), and the read actual temperature value of the temperature and the time zone, the area, and the area calculated by the air temperature prediction unit 114 A difference (prediction error) with the predicted value of the air temperature corresponding to the prediction target day is calculated (S262).

  After performing the above processing, the error analysis unit 115 calculates the average value μ of the prediction error and the standard deviation σ for each area and seasonal information (S263). In addition, about the calculation process of average value (micro | micron | mu) and standard deviation (sigma), it abbreviate | omits details here as what uses a general method. The error analysis unit 115 registers the calculated average value μ and standard deviation σ in the error data storage unit 135 in association with the season information (S 264).

  As described above, the reference curve for each region is created and registered in the reference curve storage unit 133, and the average value μ and the standard deviation σ concerning the prediction error for each region and season are calculated and are stored in the error data storage unit 135. be registered.

<Prediction process>
FIG. 33 is a diagram showing a flow of air temperature prediction processing in the future.

  The temperature prediction unit 114 reads the latest prediction target date stored in the prediction value storage unit 132 (S301), and the following processing is performed for each area of "Hiroshima", "Okayama", "Matsue", "Tottori" and "Yamaguchi" To be done.

  The air temperature prediction unit 114 reads the predicted values of the highest temperature and the lowest temperature corresponding to the prediction target day and area from the prediction value storage unit 132 (S302), and reads the reference curve corresponding to the area from the reference curve storage unit 133 (S303) ). The air temperature prediction unit 114 multiplies the normalized value of each time zone included in the read reference curve by a value obtained by subtracting the predicted value of the lowest air temperature read out from the read out predicted value of the highest air temperature, The predicted value of the temperature for each time zone is calculated by adding the read predicted value of the lowest temperature (S304). The probability distribution creation unit 116 creates a probability distribution of the calculated predicted value (S305).

  FIG. 34 is a diagram showing a flow of processing for creating a probability distribution of predicted air temperature values for each time zone.

  The probability distribution creation unit 116 identifies season information corresponding to the prediction target date (S321). As described above, the season information is determined according to a predetermined rule according to the prediction target date, but the probability distribution creation unit 116 may receive input of the season information.

  The probability distribution creating unit 116 reads the average value μ and the standard deviation σ corresponding to the season information from the error data storage unit 135 (S322), and uses the read average value μ and the standard deviation σ to calculate the prediction error for each time zone A probability density function f (e) of normal distribution according to e is created (S323).

  The probability distribution creation unit 116 obtains X percentile values for prediction errors from the probability density function f (e) for a time zone, where X is 10, 30, 50, 70, and 90, respectively (S324). The X percentile value of the predicted temperature value is calculated by adding the X percentile value of the prediction error to the predicted value of (S325).

  Referring back to FIG. 33, after performing the above-described processing for each time zone, the predicted value output unit 117 creates a predicted value of a time zone in units of 30 minutes (S306).

  FIG. 35 is a diagram showing a flow of processing for calculating a predicted value in a time zone of 30 minutes.

The predicted value output unit 117 sets the X percentile value of the predicted air temperature value related to the time zone i-1 immediately before the time zone i to X, assuming 10, 30, 50, 70, 90 as X for each time zone i. _i-1 ^X (S341), the X percentile value of the predicted air temperature value for the time zone i is T _i ^X (S342), and the X percentile value of the air temperature forecast value for the next time zone i + 1 of the time zone i ^Let T _{i + 1 be} ^X (S343), and the predicted value iA ^{X of the} air temperature related to the time zone iA for 30 minutes starting from the hour of the time zone i concerned is calculated by the following equation (S344).

Further, the predicted value output unit 117 calculates the predicted value iB ^X of the air temperature related to the time zone iB for 30 minutes starting 30 minutes after the hour of the time zone i according to the following equation (S345).

Referring back to FIG. 33, the predicted value output unit 117 outputs the probability distribution (10th, 30th, 50th, 70th, and 90th percentile values) of the predicted air temperature value in the 30-minute time zone as the predicted air temperature value (S307).

FIG. 36 is a diagram showing an example of the predicted temperature value output by the predicted value output unit 117. As shown in FIG. As shown in the figure, the predicted value output unit 117 calculates the 10% value (estimated value iA ^10% , iB ^10% ) of the predicted air temperature value for each of the time zones iA and iB for 30 minutes from 1A to 24B, 30% (predicted value iA ^30% , iB ^30% ), 50% value (predicted value iA ^50% , iB ^50% ), 70% value (predicted value iA ^70% , iB ^70% ) and 90% value (predicted value) The values iA ^90% and iB ^90% are output. As described above, the predicted value output unit 117 can display the predicted values of the air temperature in a list including variations thereof. Thus, the user can grasp the prediction of the air temperature in more detail.

=== Effect of temperature prediction ===
As described above, according to the air temperature prediction device 10 of the present embodiment, a reference curve representing a change in air temperature per day is created based on the past actual temperature value, and the air temperature is predicted based on the reference curve. It can be performed. Therefore, for example, instead of using the average value or the median value of past actual temperature values as the current temperature, it is possible to predict the temperature according to the tendency of the daily temperature change, and the prediction accuracy is improved. It is expected.

  Moreover, in the air temperature prediction apparatus 10 of this embodiment, the value normalized so that the change from the lowest temperature to the highest temperature may be used is used as a reference | standard curve. Therefore, it is possible to conduct an analysis focusing on the change of the temperature regardless of the range of the temperature difference and the rise and fall of the lowest maximum temperature. Therefore, it becomes possible to grasp the tendency of change of temperature and to utilize for temperature prediction, and the improvement of prediction accuracy is expected.

  Moreover, although the change condition of the air temperature on the 1st may differ with areas, in the air temperature prediction apparatus 10 of this embodiment, the standard curve is created for every area. Therefore, the reference curve of the present embodiment can appropriately represent the characteristic temperature change in the area, and the prediction accuracy of the air temperature can be improved.

  Moreover, although the change condition of the air temperature of 1st may differ with days, in the air temperature prediction apparatus 10 of this embodiment, the standard curve is created for every date. Therefore, the reference curve reflects characteristics such as the season and reflects the characteristic temperature change depending on the day, so that the temperature change of one day can be represented, and the prediction accuracy of the air temperature can be improved.

  Moreover, according to the air temperature prediction apparatus 10 of this embodiment, the prediction error concerning the prediction value of the air temperature by the reference curve is calculated, and the correction based on the prediction error is performed. Therefore, the prediction accuracy can be improved.

  Moreover, in the air temperature prediction apparatus 10 of this embodiment, the probability distribution of the air temperature predicted value is generated from the probability distribution of the prediction error. This makes it possible to grasp the variation in the predicted temperature value, which is useful for making decisions based on the temperature. Further, in the air temperature prediction apparatus 10 of the present embodiment, the error data is managed for each time zone, so that the prediction accuracy can be kept high corresponding to the case where the prediction error varies depending on the time zone.

  In the present embodiment, the air temperature prediction apparatus 10 is a single computer, but may be realized by a plurality of computers. For example, one virtual computer can be realized by a plurality of computers, or a system can be configured by a processing computer having a functional unit mounted thereon and a database server mounted with a storage unit. it can. A weather server may be provided to manage weather data (including forecast values by the Meteorological Agency and measurement results by measuring instruments), and the weather prediction device 10 may inquire the weather server.

  In the present embodiment, the air temperature prediction is performed only by the air temperature prediction device 10. However, the present invention is not limited to this. For example, a computer that creates a reference curve, a computer that analyzes an error, and a computer that performs prediction It can be another computer.

  Further, in the present embodiment, the reference curve is determined by averaging the prediction errors. However, the present invention is not limited to this. For example, an approximate expression may be determined by curve fitting to be used as the reference curve.

  In the present embodiment, the reference curve is created for each area and date and stored in the reference curve storage unit 133. However, the present invention is not limited to this, and one reference curve is created for each date common to each area. May be stored, or local and seasonal reference curves may be created and stored. A reference curve for each season common to each area may be created and stored.

  Further, in the present embodiment, the statistical value (average value μ and standard deviation σ) of the prediction error is obtained for each season and stored in the error data storage unit 135. However, one statistical error value as a whole is used. It may be determined and stored, or statistical values of prediction errors may be determined and stored for each area. Further, statistical values of prediction errors corresponding to the region and the season may be obtained and stored.

=== Effect of Demand Forecasting ===
As described above, the air temperature prediction unit 102 of the prediction device 100 acquires the predicted value of the air temperature predicted by the air temperature prediction device 10, and the demand prediction unit 103 can predict the amount of power demand using this. As described above, according to the air temperature prediction device 10 of the present embodiment, since the improvement of the prediction accuracy is expected and the amount of power demand is greatly affected by the air conditioning demand due to the air temperature, An improvement in prediction accuracy is also expected.

  In the present embodiment, although the air temperature prediction apparatus 10 is a computer different from the prediction apparatus 100, the prediction apparatus 100 includes the functional unit and the storage unit of the air temperature prediction apparatus 10, The prediction method of the air temperature according to the present embodiment using the reference curve may be used.

=== Conclusion ====
As mentioned above, said each embodiment can be described as follows.

  Each of the above embodiments is a prediction system that predicts a profit value (Y) in power transaction, and a first storage unit that stores weather information of a plurality of days in a time zone P in which the power transaction is performed (the above embodiment In the second storage unit (which corresponds to the data table M6) and the data representing the relationship between the amount of power demand and the power generation cost (in the above embodiment, the data relating to the marginal cost line calculated by the data tables M1 and M2) , And a third storage unit (corresponding to the data table M3 in the above embodiment) that stores the power transaction amount in the time zone P and the transaction price, and data representing the relationship between the temperature and the power demand amount. A fourth storage unit (in the above embodiment, corresponding to Equations (13), (15), (17), (18), (19) in the above embodiment)) and a plurality of days in time zone P Based on the weather information, time zone P Based on the temperature prediction unit (102) that calculates the probability distribution data of temperature, the probability distribution data of predicted temperature, and the data representing the relationship between the temperature and the amount of power demand, the amount of power demand in time zone P (N Calculated from the demand forecasting unit (103, 103 ') which calculates probability distribution data of), probability distribution data of power demand amount, data correlating the power demand amount with generation cost, and the transaction amount (C) in the power transaction Calculating the probability distribution data of profit value in the power transaction based on the difference between the change cost of the power generation cost in the case of conducting the power transaction of the transaction volume and the transaction price in the case of conducting the power transaction of the transaction volume And a profit prediction unit (104).

  By this, the power supplier can recognize the probability distribution (for example, fluctuation range) of the profit amount when the power transaction is performed.

  Here, the prediction system has a fifth storage unit (corresponding to the data table M4 in the above embodiment) storing the past data related to the measured air temperature stored in association with the time at which the measurement was performed. It further includes a sixth storage unit (corresponding to the data table M7 in the above embodiment) that stores past data related to the amount of power demand stored in association with time, and the demand prediction unit And the regression equation of the temperature and the power demand based on the past data on the power demand (in the above embodiment, the equations (13), (15), (17), (18), and (18), and (19) ) May be calculated, and the probability distribution data of the power demand amount of the time zone P may be calculated based on the regression equation and the probability distribution data of the predicted air temperature. At this time, the demand prediction unit calculates a regression equation using at least a difference between the predicted air temperature and about 18 degrees Celsius and a square of the difference between the predicted air temperature and about 18 degrees Celsius as an explanatory variable. In addition, the prediction system further includes a seventh storage unit (corresponding to the data table M9 in the above embodiment) that stores the prediction reference value of the power demand in the time zone P, and the regression equation represents the probability distribution of predicted air temperature. It may be a regression equation (corresponding to the equation (17), the equation (18), and the equation (19) in the above embodiment) representing the amount of fluctuation from the predicted reference value of the power demand based on the data.

  By this, the power supplier can accurately reflect the probability distribution of the temperature on the probability distribution of the power demand amount.

  Here, the profit prediction unit may calculate the amount of traded power at which the profit value is maximum, based on the probability distribution data of the profit value.

  By this, the power supplier can maximize profit.

  Here, the data representing the relationship between the power demand amount and the power generation cost stored in the second storage unit includes the power generation potential of the plurality of power generation means, the power generation cost per unit power generation amount of the plurality of power generation means, It may be data on a marginal cost line calculated based on the operation priorities for a plurality of power generation means.

  Here, the data representing the relationship between the power transaction amount in the time zone P and the transaction price stored in the third storage unit is a transaction price corresponding to the power transaction amount based on the transaction price per unit power generation amount. It may be data to be calculated.

  Further, each of the above embodiments is a prediction method for predicting a profit value in power trading, and calculates the temperature distribution of predicted temperature of time zone P based on weather information of a plurality of days of time zone P. A demand forecasting process for calculating probability distribution data of the power demand in the time zone P based on the forecasting process, the probability distribution data of the predicted air temperature, and the data representing the relationship between the air temperature and the power demand, Of the power generation cost in the case of conducting the power transaction of the above-mentioned transaction volume calculated from the probability distribution data of the data, the data of correlating the power demand amount and the power generation cost, and the transaction volume in the power transaction And a profit prediction step of calculating probability distribution data of a profit value in a power transaction based on the difference between the price and the transaction price in the case of performing the transaction.

  By this, the power supplier can recognize the probability distribution (for example, fluctuation range) of the profit amount when the power transaction is performed.

  Although the specific examples of the present invention have been described above in detail, these are merely examples and do not limit the scope of the claims. The art set forth in the claims includes various variations and modifications of the specific examples illustrated above.

DESCRIPTION OF SYMBOLS 100 forecasting apparatus 101 acquisition part 102 air temperature forecasting part 103 demand forecasting part 104 profit forecasting part 105 presentation part 200 transaction information provision apparatus 300 communication network 400 weather information provision apparatus 500 demand information provision apparatus P time zone of prediction object

Claims (7)

A system that predicts the demand for electricity,
A first storage unit that stores weather information;
A third storage unit storing data representing the relationship between the temperature and the demand amount;
A fourth storage unit storing data representing the relationship between sunshine hours and the demand amount;
A fifth storage unit for storing past data on measured temperature;
A sixth storage unit for storing past data on the amount of power demand;
A seventh storage unit for storing past data regarding sunshine hours;
A temperature prediction unit that calculates probability distribution data of predicted temperature based on the weather information;
A sunshine time prediction unit for acquiring the predicted value of the sunshine time;
Based on the data representing the relationship between the temperature and the demand amount, the data representing the relationship between the sunshine duration and the demand amount, the probability distribution data of the predicted temperature, and the predicted value of the sunshine duration, the probability distribution data of the demand amount Demand forecasting part to calculate,
Equipped with
The demand forecasting unit is based on past data on the measured air temperature, past data on the demand amount, and past data on the sunshine duration, and at least a difference of about 18 degrees Celsius from the predicted air temperature; Using the square of the difference between the predicted temperature and approximately 18 degrees Celsius and the sunshine time as an explanatory variable, calculate the regression equation of the temperature and sunshine time and the demand amount, and calculate the regression equation and the probability distribution data of the predicted temperature and the above based on the predicted value of the daylight hours, the demand of the probability distribution data electricity demand forecasting system, characterized that you calculated. The power demand forecasting system according to claim 1, wherein
The sunshine time prediction unit acquires probability distribution data of a predicted value of the sunshine time as a predicted value of the sunshine time.
Power demand forecasting system characterized by A system that predicts the demand for electricity,
An actual value storage unit that stores actual values of temperature for each time zone;
A reference curve creation unit that normalizes the actual value and creates a reference curve that represents a change in temperature for each time zone;
A first storage unit that stores weather information;
A third storage unit storing data representing the relationship between the temperature and the demand amount;
A fourth storage unit storing data representing the relationship between sunshine hours and the demand amount;
A temperature prediction unit that calculates probability distribution data of predicted temperatures based on the weather information, and calculates predicted temperatures for each of the time zones based on the maximum temperature and the minimum temperature and the reference curve ;
A maximum minimum temperature acquisition unit for acquiring past predicted values and future prediction values of the maximum temperature and the minimum temperature;
A prediction error calculation unit which gives the predicted temperature of the past to the temperature prediction unit by giving the predicted values of the past of the maximum temperature and the minimum temperature, and calculates a prediction error of the predicted temperature of the past;
A sunshine time prediction unit for acquiring the predicted value of the sunshine time;
Based on the data representing the relationship between the temperature and the demand amount, the data representing the relationship between the sunshine duration and the demand amount, the probability distribution data of the predicted temperature, and the predicted value of the sunshine duration, the probability distribution data of the demand amount Demand forecasting part to calculate,
Equipped with
The air temperature prediction unit calculates the future predicted air temperature based on the highest temperature, the lowest air temperature, and the reference curve, and based on the future predicted air temperature and the prediction error, probability distribution data of the predicted air temperature electricity demand forecasting system, characterized that you calculated. A method of forecasting the amount of electricity demand,
The computer is
Storing weather information;
Storing data representing the relationship between the temperature and the demand amount;
Storing data representing a relationship between daylight hours and the demand amount;
Storing past data on measured temperature;
Storing historical data on the demand for electricity;
Storing past data about daylight hours;
Calculating probability distribution data of predicted air temperature based on the weather information;
Obtaining a predicted value of the sunshine duration;
Based on the data representing the relationship between the temperature and the demand amount, the data representing the relationship between the sunshine duration and the demand amount, the probability distribution data of the predicted temperature, and the predicted value of the sunshine duration, the probability distribution data of the demand amount Calculating steps;
Run
The step of calculating the probability distribution data of the demand amount is based on at least the predicted air temperature on the basis of past data on the measured air temperature, past data on the demand amount, and past data on the sunshine duration. Using the difference of 18 degrees Celsius, the square of the difference between the predicted air temperature and the difference of about 18 degrees Celsius, and the sunshine hours as the explanatory variables, a regression equation of the temperature and sunshine hours and the demand amount is calculated. A method of predicting a power demand amount , wherein the probability distribution data of the demand amount is calculated based on the probability distribution data of predicted air temperature and the predicted value of the sunshine duration . It is a program for forecasting the demand for electricity,
On the computer
Storing weather information;
Storing data representing the relationship between the temperature and the demand amount;
Storing data representing a relationship between daylight hours and the demand amount;
Storing past data on measured temperature;
Storing historical data on the demand for electricity;
Storing past data about daylight hours;
Calculating probability distribution data of predicted air temperature based on the weather information;
Obtaining a predicted value of the sunshine duration;
Based on the data representing the relationship between the temperature and the demand amount, the data representing the relationship between the sunshine duration and the demand amount, the probability distribution data of the predicted temperature, and the predicted value of the sunshine duration, the probability distribution data of the demand amount Calculating steps;
A program for executing,
The step of calculating the probability distribution data of the demand amount is based on at least the predicted air temperature on the basis of past data on the measured air temperature, past data on the demand amount, and past data on the sunshine duration. Using the difference of 18 degrees Celsius, the square of the difference between the predicted air temperature and the difference of about 18 degrees Celsius, and the sunshine hours as the explanatory variables, a regression equation of the temperature and sunshine hours and the demand amount is calculated. The probability distribution data of the demand amount is calculated based on the probability distribution data of the predicted temperature and the predicted value of the sunshine time
A program characterized by A method of forecasting the amount of electricity demand,
The computer is
Storing the actual temperature value for each time zone;
Normalizing the actual value to create a reference curve representing a change in temperature for each time zone;
Storing weather information;
Storing data representing the relationship between the temperature and the demand amount;
Storing data representing a relationship between daylight hours and the demand amount;
Calculating probability distribution data of predicted air temperature based on the meteorological information, and calculating predicted air temperature for each of the time zones based on a maximum temperature and a minimum temperature and the reference curve;
Obtaining respective past forecast values and future forecast values of the maximum temperature and the minimum temperature;
A step of the highest temperature and on the basis of the past estimated value and the reference curve of the minimum temperature to calculate the past the prediction temperature of each of the time zones, and calculates a prediction error of the prediction temperatures of the past,
Obtaining a predicted value of the sunshine duration;
Based on the data representing the relationship between the temperature and the demand amount, the data representing the relationship between the sunshine duration and the demand amount, the probability distribution data of the predicted temperature, and the predicted value of the sunshine duration, the probability distribution data of the demand amount Calculating steps;
Run
The step of calculating the predicted air temperature for each of the time zones calculates the future predicted air temperature based on the maximum temperature and the minimum air temperature and the reference curve, and based on the future predicted air temperature and the prediction error And calculating a probability distribution data of the predicted air temperature. It is a program for forecasting the demand for electricity,
On the computer
Storing the actual temperature value for each time zone;
Normalizing the actual value to create a reference curve representing a change in temperature for each time zone;
Storing weather information;
Storing data representing the relationship between the temperature and the demand amount;
Storing data representing a relationship between daylight hours and the demand amount;
Calculating probability distribution data of predicted air temperature based on the meteorological information, and calculating predicted air temperature for each of the time zones based on a maximum temperature and a minimum temperature and the reference curve;
Obtaining respective past forecast values and future forecast values of the maximum temperature and the minimum temperature;
A step of the highest temperature and on the basis of the past estimated value and the reference curve of the minimum temperature to calculate the past the prediction temperature of each of the time zones, and calculates a prediction error of the prediction temperatures of the past,
Obtaining a predicted value of the sunshine duration;
Based on the data representing the relationship between the temperature and the demand amount, the data representing the relationship between the sunshine duration and the demand amount, the probability distribution data of the predicted temperature, and the predicted value of the sunshine duration, the probability distribution data of the demand amount Calculating steps;
A program to execute
The step of calculating the predicted air temperature for each of the time zones calculates the future predicted air temperature based on the maximum temperature and the minimum air temperature and the reference curve, and based on the future predicted air temperature and the prediction error A program for calculating probability distribution data of the predicted temperature.