JP2009251742A - Demand predicting device, demand predicting method and demand predicting program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To accurately predict power demand by intelligible logic. <P>SOLUTION: A data incorporating part 110 incorporates the amount demanded at every fixed time, calendar data and weather forecast and stores them in an incorporated data storage part 120. An average amount demanded computing part 130 computes the average amount demanded of the same time in the past using data in the incorporated data storage part 120. A deviation computing part 140 computes deviation between the average amount demanded and the actual amount demanded. A predicting part 160 estimates a prediction expression based on the average amount demanded and the deviation and predicts the amount demanded. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a demand forecasting device, a demand forecasting method, and a demand forecasting program that forecast future demand based on past demand, and in particular, demand that can be predicted with easy-to-understand logic while maintaining high accuracy. The present invention relates to a prediction device, a demand prediction method, and a demand prediction program.

  In order to stably supply power and heat, it is necessary to produce power and heat that meet demand. However, for example, when excessive power generation is prepared in the power supply, inefficient generator operation is performed and the economy is impaired. For this reason, preparation of power generation is performed based on demand prediction.

  Predicting demand with high accuracy is extremely important for economical generator operation, and many demand prediction technologies have been developed. As such a technique, for example, there are techniques described in Patent Documents 1 to 4 and the like.

JP 2007-215354 A JP 2005-328673 A JP 2005-224002 A JP 2005-168251 A

  However, in order to make a practical prediction device, it is desirable that an operator who actually prepares a power generation preparation plan based on demand prediction has an easy-to-understand method and logic for predicting demand. . For this reason, as a practical demand amount prediction technique, it is desired to develop a technique for performing prediction with easy-to-understand logic while maintaining high prediction accuracy.

  The present invention has been made to solve the above-described problems caused by the prior art, and is a demand prediction device, a demand prediction method, and a demand prediction program capable of performing prediction with easy-to-understand logic while maintaining high prediction accuracy. The purpose is to provide.

  In order to solve the above-described problems and achieve the object, the present invention is a demand prediction device that predicts a future demand amount based on a past demand amount, and is an average demand at each time that is a prediction target of the demand amount. An average demand amount calculating means for calculating the amount, a deviation calculating means for calculating a deviation at the current time based on a deviation between the average demand amount calculated by the average demand amount calculating means and the actual demand amount, and a prediction target time Predicting demand amount calculating means for calculating the demand amount at the prediction target time based on the average demand amount calculated by the average demand amount calculating means and the deviation calculated by the deviation calculating means with respect to the current time. It is characterized by having.

  The present invention is also a demand prediction method by a demand prediction device that predicts a future demand amount based on a past demand amount, and calculates an average demand amount at each time that is a prediction target of the demand amount. A calculating step; a deviation calculating step for calculating a deviation at the current time based on a deviation between the average demand calculated by the average demand calculating step and the actual demand; and the average demand for the prediction target time A predicted demand amount calculating step of calculating a demand amount at a prediction target time based on the average demand amount calculated by the calculating step and the deviation calculated from the current time by the deviation calculating step. .

  Further, the present invention is a demand prediction program for predicting a future demand amount based on a past demand amount, and an average demand amount calculation procedure for calculating an average demand amount at each time that is a target of the demand amount prediction; Deviation calculation procedure for calculating the deviation at the current time based on the deviation between the average demand calculated by the average demand calculation procedure and the actual demand, and the average demand calculation procedure for the prediction target time The computer is caused to execute a predicted demand amount calculation procedure for calculating a demand amount at a prediction target time based on the calculated average demand amount and the deviation calculated with respect to the current time by the deviation calculation procedure.

  According to this invention, the average demand amount at each time that is the target of the demand amount calculation is calculated, the deviation at the current time is calculated based on the deviation between the calculated average demand amount and the actual demand amount, and the prediction target time Since the demand amount at the prediction target time is calculated based on the average demand amount calculated with respect to the current time and the deviation calculated with respect to the current time, the demand amount can be accurately calculated with easy-to-understand logic.

  According to the present invention, since the demand amount is accurately calculated with easy-to-understand logic, there is an effect that the demand prediction can be put into practical use.

  Exemplary embodiments of a demand prediction apparatus and a demand prediction method according to the present invention will be described below in detail with reference to the accompanying drawings.

  First, the concept of power demand prediction according to the present embodiment will be described. 1 and 2 are explanatory diagrams for explaining the concept of power demand prediction according to the present embodiment. Fig. 1 is a conceptual diagram showing the amount of demand for each time in black and white for one year. Fig. 2 shows the deviation between the average demand for each time and the actual demand for one year in black and white. FIG.

  As shown in FIG. 1, power demand is regular with respect to time throughout the year. That is, if the time is the same, the same amount of demand tends to be the same even if the dates are different. Therefore, when power demand prediction is performed, it is possible to predict with high accuracy by performing prediction based on time.

  Therefore, it is first considered to make a prediction based on the average demand at the prediction target time. Then, as shown in FIG. 2, the deviation between the average demand and the actual demand has regularity with respect to the date. That is, on the same day, the deviation tends to be the same even if the time is different.

  Thus, the power demand has regularity with respect to time throughout the year, and the deviation between the average demand and the actual demand has regularity with respect to the date. Therefore, demand prediction can be performed with high accuracy by making demand prediction using these regularities.

  Therefore, the power demand prediction apparatus according to the present embodiment predicts the power demand based on the average demand amount at the prediction target time and the deviation of the prediction target date. Thus, by predicting the power demand based on the average demand amount of the prediction target time and the deviation of the prediction target date, the power demand prediction device according to the present embodiment is easy to understand while maintaining high prediction accuracy. Predictions can be made.

  Next, the configuration of the power demand prediction apparatus according to the present embodiment will be described. FIG. 3 is a functional block diagram illustrating the configuration of the power demand prediction apparatus according to the present embodiment. As shown in the figure, the power demand prediction apparatus 100 includes a data capture unit 110, a captured data storage unit 120, an average demand amount calculation unit 130, a deviation calculation unit 140, and a calculated data storage unit 150. And a prediction unit 160.

  The data fetching unit 110 includes demand data for each fixed time, calendar data, that is, calendar data such as day of the week, whether it is a public holiday, whether it is a New Year period, whether it is a GW period, whether it is a Bon Festival period, weather forecast data, etc. Is a processing unit that captures data and writes it into the captured data storage unit 120. The captured data storage unit 120 is a storage unit that stores data captured by the data capturing unit 110.

  FIG. 4 is a diagram illustrating an example of the captured data storage unit 120. As shown in the figure, the captured data storage unit 120 includes the latest power demand for a predetermined number of days every minute, date, time, day of the week, whether it is a national holiday, whether it is a New Year period, a GW period And whether it is a Bon period, temperature forecast, and weather forecast information.

The average demand amount calculation unit 130 is a processing unit that reads the data stored in the captured data storage unit 120 and calculates the average demand amount. Assuming that the current time is time t on day d, the average demand amount calculation unit 130 calculates n days (n = d−p) from the demand amount stored in the captured data storage unit 120 up to the previous day according to Equation (1). , d−p + 1,..., d−1, d), the average demand amount avD _{n, t} for the past m days at time _t is calculated.

Here, p is a parameter that represents a predetermined number of days used for estimating the prediction formula, and m is a parameter that represents a predetermined period for calculating the average demand. In Expression (1), D _{n, t} represents the demand amount at time t on n days stored in the captured data storage unit 120.

Similarly, the average demand amount calculation unit 130 calculates the average demand amount avD _{n, t} at time t + τ of n days (n = d−p, d−p + 1,..., D−1, d) in Expression (2). Calculate _{+ τ} . Here, τ represents a time interval for performing prediction.

In addition, here, the average demand amount avD _{n, t} is calculated by a simple average as in Expression (1), but it can also be calculated by Expression (3) in which past data is weighted.

In formula (3), w _i represents the weight of the demand amount for n−i days. As the weight, for example, there is a method such as increasing the weight as the nearest day, or increasing the weight of a day with a high degree of similarity such as the predicted day and the day of the week being the same. The average demand amount calculation unit 130 writes the calculated average demand amount in the calculation data storage unit 150 in association with the date.

  The deviation calculation unit 140 is a processing unit that calculates a deviation between the average demand calculated by the average demand calculation unit 130 and the actual demand. That is, the deviation calculating unit 140 calculates a deviation of how far the actual power demand has deviated from the average demand calculated by the average demand calculating unit 130 in the past. The deviation calculating unit 140 writes the calculated deviation in the calculated data storage unit 150 in association with the date.

Deviation Δ _{n, t} from the average demand at time t on n days (n = d−p, d−p + 1,..., d−1, d) is calculated as an exponential smoothing value _expressed by equation (4). Is done.

Here, r is a real number between 0 and 1. If r = 1.0, Δ _{n, t} is a simple deviation at time t. If r <1.0, Δ _{n, t} is not a simple deviation at the time t, but a deviation reflecting a trend of deviation before the time t.

The calculated data storage unit 150 is a storage unit that stores data calculated by the average demand calculation unit 130, the deviation calculation unit 140, and the prediction unit 160. FIG. 5 is a diagram illustrating an example of the calculated data storage unit 150. As shown in the figure, the calculated data storage unit 150 stores the average demands avD _{n, t} and avD _{n, t + τ} and the deviation Δ _{n, t} for (p + 1) days. Note that Sat _n and Sun _n are calculated by the prediction unit 160.

  The predicting unit 160 uses the average demand calculated by the average demand calculating unit 130, the deviation calculated by the deviation calculating unit 140, the calendar data captured by the data capturing unit 110, and the like as variables. Is a processing unit that estimates a future demand amount using the prediction formula.

Specifically, the prediction unit 160 first calculates a day-of-week dummy variable for reflecting the difference in day of the week in the demand prediction from the calendar data stored in the captured data storage unit 120. For example, demand is different on Saturdays and Sundays on weekdays. In order to reflect this difference in the prediction, the prediction unit 160 calculates a Saturday dummy variable Sat _n and a Sunday dummy variable Sun _n for the calendar data for the past p + 1 days. Here, n = d, d−1,..., D−p.

  The Saturday dummy variable is a variable that takes a value of 1 if n day is Saturday, and 0 otherwise, and the Sunday dummy variable is 1 if n day is Sunday, Golden Week period, Bon Festival / New Year period. In other cases, the variable takes a value of 0. It should be noted that other calculation methods can be used as the calculation method of the day-of-week dummy variable.

Then, the prediction unit 160 reads the average demand amount avD _{n, t + τ} and the deviation Δ _{n, t} stored in the calculated data storage unit 150. Here, n = d-1, d-2, ..., dp. Then, the prediction unit 160 reads the value of the power demand amount D _{n, t + τ from} the captured data storage unit 120, and uses the read data of the past p days (d−1 days to d−p days). Estimate the prediction formula.

As a method for estimating the prediction formula, application of multiple regression analysis, neural network, or the like can be considered. Here, the prediction formula is estimated by multiple regression analysis. As the estimation of the prediction formula by the multiple regression analysis, the values of the coefficients a ₁ , a ₂ , a ₃ , a ₄ and a ₅ of the prediction formula that minimize the value of E calculated by the formula (5) are obtained.

Then, the prediction unit 160 finally calculates the power demand amount at time t + τ using Expression (6) using the obtained prediction expression, and displays the calculation result on the display device.

In addition to the calendar data, a predicted temperature value can be added as a variable of the prediction formula. For example, with regard to the temperature, a cooling device is used when the temperature exceeds a certain value, and a heating device is used when the temperature falls below a certain value. In order to reflect the demand amount of these air conditioning equipment, for example, the variable _Td calculated by the equation (7) is added to the variable of the prediction equation.

Here, C and H are predetermined parameters such that C ≧ H, C corresponds to the temperature at which the cooling device is used, and H corresponds to the temperature at which the heating device is used. FT _d is the expected temperature. In this case as well, a prediction formula (8) that minimizes the square error is obtained for the data from the past d−p to d−1, and the demand at time t + τ is predicted using this formula.

Further, in order to reflect the change in demand due to the lighting load in the prediction of the demand amount, a weather dummy variable W _d can be added as a variable of the prediction formula. For example, the weather dummy variable W _d can be calculated as shown in Equation (9), and the prediction equation can be changed as shown in Equation (10).

  Next, a processing procedure of demand prediction processing by the power demand prediction apparatus 100 according to the present embodiment will be described. FIG. 6 is a flowchart illustrating a processing procedure of a demand prediction process performed by the power demand prediction apparatus 100 according to the present embodiment. Here, it is assumed that the data capturing unit 110 constantly captures the latest data such as the power demand and writes it in the captured data storage unit 120.

As shown in FIG. 6, in the power demand prediction apparatus 100, the average demand calculation unit 130 calculates the average demands avD _{n, t} and avD _{n, t + τ} for (p + 1) days at the current time t and the prediction target time t + τ. And is written in the calculated data storage unit 150 (step S1).

Then, the deviation calculating unit 140 calculates the deviation Δ _{n, t} for (p + 1) days at the current time t, and writes it in the calculated data storage unit 150 (step S2). Further, the prediction unit 160 calculates the values of the Saturday dummy variable and the Sunday dummy variable for p days, and writes them in the calculated data storage unit 150.

Then, the prediction unit 160 uses the demand amount D _{n, t + τ} , the average demand amount avD _{n, t + τ} , the deviation Δ _{n, t} , the Saturday dummy variable value, and the Sunday dummy variable value for p days. Estimation (step S3), and using the estimated prediction formula, a predicted demand amount at the prediction target time t + τ is calculated (step S4).

  As described above, the power demand prediction apparatus 100 can accurately predict the demand amount with easy-to-understand logic by calculating the predicted demand amount based on the average demand amount at the prediction target time and the deviation at the current time.

  Next, the processing procedure of the average demand calculation process by the average demand calculation part 130 is demonstrated. FIG. 7 is a flowchart showing a processing procedure of an average demand amount calculation process by the average demand amount calculation unit 130. As shown in the figure, the average demand amount calculation unit 130 initializes a variable n representing a date for calculating the average demand amount on the prediction target date. That is, n = d is set (step S11a).

Then, the variable i for counting the number of days is initialized to 1, and the total value Ave of the demand amount is initialized to 0 (step S11b). Then, the demand amount D _{ni, t} of date _{ni at time t} is read from the captured data storage unit 120 (step S11c) and added to Ave (step S11d).

Then, i is increased by 1 (step S11e), and it is determined whether i is larger than the average value calculation days m (step S11f). As a result, if i is not larger than the average value calculation days m, the process returns to step S11c. If i is larger than the average value calculation days m, the average at date n and time t is obtained by dividing Ave by m. The demand amount avD _{n, t} is calculated (step S11g).

Then, the calculated average demand amount avD _{n, t} is written in the calculated data storage unit 150 in correspondence with the date (step S11h), and n is decreased by 1 (step S11i). Then, it is determined whether or not the average demand for (p + 1) days has been calculated, that is, whether or not n is smaller than d−p (step S11j), and when n is not smaller than d−p, Returning to step S11b, if n is smaller than d−p, the average demand calculation process is terminated.

Thus, the average demand amount calculation unit 130 can calculate the average demand amount using the past demand amount at the same time. Here, the date n, the average demand AVD n at time _t, has been described for calculating the _t, demand D _ni, demand instead of _t amount D _ni, by using the _{t + tau,} date n, The average demand amount avD _{n, t + τ} at time t + τ can be calculated.

Further, here, the case of calculating the simple average has been described, but as shown in FIG. 8, the demand value is weighted to calculate the total value Ave and the sum S of the weights w _i (1 ≦ i ≦ m) is calculated. The weighted average demand can be calculated by Ave / S.

  Next, a processing procedure of deviation calculation processing by the deviation calculation unit 140 will be described. FIG. 9 is a flowchart showing a procedure of deviation calculation processing by the deviation calculation unit 140. As shown in the figure, the deviation calculating unit 140 initializes a variable n representing a date for calculating the deviation to dp (step S21a), and initializes a variable i for counting time to 0 (step S21b).

Then, the demand amount D _{n, i} at date n at time i is read from the captured data storage unit 120 (step S21c), and the average demand amount avD _{n, i} at date n and time _i is calculated (step S21d). Then, the deviation Δ _{n, i} at date n and time _i is calculated (step S21e). The deviation Δ _{n, −τ} is assumed to be zero.

The deviation Δ _{n, i} is temporarily stored for use in calculating the deviation Δ _{n, i + 1} (step S21f), and i is increased by τ (step S21g). Then, it is determined whether i is larger than t (step S21h). If i is not larger than t, the process returns to step S21c to calculate the deviation Δ _{n, i} for the next i.

On the other hand, when i is larger than t, the deviation Δ _{n, i,} that is, the deviation Δ _{n, t} is stored in the calculated data storage unit 150 (step S21i), and n is incremented by 1 (step S21j). Then, it is determined whether or not n is larger than d (step S21k). If not larger than d, the process returns to step S21b to calculate a deviation Δ _{n, t} for the next n. On the other hand, if n is greater than d, the deviation calculation process is terminated.

In this way, the deviation calculating unit 140 calculates the deviation Δ _{n, t} based not only on the difference between the average demand at the current time and the actual demand, but also on the deviation up to the current time on the current day, The trend of deviation on the day can be reflected.

  Next, a processing procedure of prediction formula estimation processing by the prediction unit 160 will be described. FIG. 10 is a flowchart illustrating a processing procedure of prediction formula estimation processing by the prediction unit 160. As shown in the figure, the prediction unit 160 initializes a variable n representing the date of data used for calculation of the prediction formula to d−1 (step S31a), and obtains calendar data for n days from the captured data storage unit 120. Read (step S31b).

Then, the value of the day-of-week dummy variable is calculated from the read calendar data (step S31c), and the average demand avD _{n, t + τ} and the deviation Δ _{n, t} stored in the calculated data storage unit 150 are read (steps S31d to S31d). Step S31e). Further, the demand amount D _{n, t + τ} is read from the captured data storage unit 120 (step S31f).

Then, n is decremented by 1 (step S31g), and it is determined whether or not n is equal to or greater than d−p, that is, whether or not data for p days used for estimation of the prediction formula is still insufficient (step S31h). . As a result, when n is equal to or greater than d−p, there is not enough data, so the process returns to step S31b to align the next data. On the other hand, when n is not equal to or greater than d−p, data for p days has been prepared. Therefore, an equation for predicting the demand amount D _{d, t + τ} at time t + τ from the data for p days is used by multiple regression analysis. (Step S31i).

Thus, by estimating the prediction formula using the average demand amount avD _{n, t + τ} , the deviation Δ _{n, t} , and the day-of-week dummy variable, it is possible to perform highly accurate prediction with easy-to-understand logic.

FIG. 11 is a diagram illustrating a prediction result by the power demand prediction apparatus according to the present embodiment. As shown in the figure, it can be seen that the predicted demand is in good agreement with the actual demand. Here, the demand ahead of one hour is predicted using the average demand amount avD _{n, t + τ} , the deviation Δ _{n, t} , and the day of week dummy variable.

As the data, the annual demand data recorded for every hour of the East30 model was used. Moreover, the average demand amount calculation part 130 calculated the average demand amount of each time using the data for the past seven days by Formula (11).

Moreover, the deviation calculation part 140 calculated the deviation from the average demand in each time by Formula (12).

The prediction unit 160 uses the data for the past 42 days when estimating the prediction formula of the power demand amount by multiple regression analysis, and the coefficient a _{1 of the} prediction expression that minimizes E in Expression (13), a ₂ , a ₃ , a ₄ and a ₅ were determined.

  As described above, in this embodiment, the average demand calculation unit 130 calculates the average demand at the same time in the past, and the deviation calculation unit 140 calculates the deviation between the average demand and the actual demand, Since the prediction unit 160 estimates the demand amount by estimating the prediction formula based on the average demand amount and the deviation, the demand amount can be accurately estimated with easy-to-understand logic.

  Further, in this embodiment, since the prediction formula is estimated by adding the day-of-week dummy variable, the temperature forecast, the weather forecast, etc., the demand amount can be estimated with higher accuracy. Further, in the present embodiment, the deviation calculating unit 140 calculates the deviation using exponential smoothing, so that the deviation can be calculated with higher accuracy.

  In addition, although the present Example demonstrated the power demand prediction apparatus, the power demand prediction program which has the same function can be obtained by implement | achieving the structure which a power demand prediction apparatus has with software. A computer that executes this power demand prediction program will be described.

  FIG. 12 is a functional block diagram illustrating a configuration of a computer that executes a power demand prediction program according to the present embodiment. As shown in the figure, the computer 200 includes a RAM 210, a CPU 220, an HDD 230, a LAN interface 240, an input / output interface 250, and a DVD drive 260.

  The RAM 210 is a memory that stores a program and a program execution result, and the CPU 220 is a central processing unit that reads the program from the RAM 210 and executes the program. The HDD 230 is a disk device that stores programs and data, and the LAN interface 240 is an interface for connecting the computer 200 to other computers via the LAN. The input / output interface 250 is an interface for connecting an input device such as a mouse or a keyboard and a display device, and the DVD drive 260 is a device for reading / writing a DVD.

  The power demand prediction program 211 executed in the computer 200 is stored in the DVD, read from the DVD by the DVD drive 260, and installed in the computer 200. Alternatively, the power demand prediction program 211 is stored in a database or the like of another computer system connected via the LAN interface 240, read from these databases, and installed in the computer 200. The installed power demand prediction program 211 is stored in the HDD 230, read into the RAM 210, and executed by the CPU 220.

  Moreover, although the present Example demonstrated the case where electric power demand was estimated, this invention is not limited to this, For example, when applying other demands, such as a district heat demand, it is applicable similarly. it can.

  As described above, the demand prediction apparatus, the demand prediction method, and the demand prediction program according to the present invention are useful for the stable supply of electric power, and are particularly suitable when it is important to efficiently operate the generator. Yes.

It is the image figure which represented the magnitude | size of the demand amount for every time on the monochrome level for one year. It is the image figure which represented the deviation of the average demand amount for every time, and the actual demand amount for one year by the black-and-white level. It is a functional block diagram which shows the structure of the power demand prediction apparatus which concerns on a present Example. It is a figure which shows an example of an acquisition data memory | storage part. It is a figure which shows an example of a calculation data storage part. It is a flowchart which shows the process sequence of the demand prediction process by the electric power demand prediction apparatus which concerns on a present Example. It is a flowchart which shows the process sequence of the average demand amount calculation process by an average demand amount calculation part. It is a flowchart which shows the process sequence of the average demand amount calculation process by a weighted average. It is a flowchart which shows the process sequence of the deviation calculation process by a deviation calculation part. It is a flowchart which shows the process sequence of the prediction formula estimation process by a estimation part. It is a figure which shows the prediction result by the electric power demand prediction apparatus which concerns on a present Example. It is a functional block diagram which shows the structure of the computer which performs the electric power demand prediction program which concerns on a present Example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Electric power demand prediction apparatus 110 Data acquisition part 120 Acquisition data storage part 130 Average demand amount calculation part 140 Deviation calculation part 150 Calculation data storage part 160 Prediction part 200 Computer 210 RAM
211 Power demand prediction program 220 CPU
230 HDD
240 LAN interface 250 I / O interface 260 DVD drive

Claims (9)

A demand prediction device that predicts future demand based on past demand,
An average demand amount calculating means for calculating an average demand amount at each time for which the demand amount is predicted;
Deviation calculating means for calculating a deviation at the current time based on a deviation between the average demand calculated by the average demand calculating means and the actual demand;
Predicted demand amount calculation for calculating the demand amount at the prediction target time based on the average demand amount calculated by the average demand amount calculating means with respect to the prediction target time and the deviation calculated by the deviation calculating means with respect to the current time And a demand forecasting device.   The demand forecasting device according to claim 1, wherein the forecast demand quantity calculating means further calculates the demand quantity based on whether the forecast target day is a weekday, a Saturday, or a holiday.   The demand prediction apparatus according to claim 1, wherein the predicted demand amount calculation unit further calculates a demand amount based on an expected temperature at a prediction target time.   The demand prediction apparatus according to claim 1, 2 or 3, wherein the predicted demand amount calculation means further calculates a demand amount based on a weather forecast at a prediction target time.   The deviation calculating means calculates an exponential smoothing value of a deviation between an average demand calculated by the average demand calculating means and an actual demand with respect to a current time, and a deviation of a time before a unit time from the current time. The demand prediction apparatus according to claim 1, wherein the demand prediction apparatus calculates the deviation in   6. The predicted demand amount calculation means estimates a demand amount prediction formula by multiple regression analysis, and calculates the demand amount at a prediction target time using the estimated prediction formula. The demand prediction apparatus as described in one.   The demand prediction device according to claim 1, wherein the demand amount is a power demand amount. A demand prediction method using a demand prediction device that predicts future demand based on past demand,
An average demand amount calculating step for calculating an average demand amount at each time for which the demand amount is predicted;
A deviation calculating step of calculating a deviation at the current time based on a deviation between the average demand calculated by the average demand calculating step and the actual demand;
Predicted demand amount calculation for calculating the demand amount at the prediction target time based on the average demand amount calculated by the average demand amount calculation step with respect to the prediction target time and the deviation calculated with respect to the current time by the deviation calculation step A demand forecasting method characterized by including steps. A demand forecasting program that predicts future demand based on past demand,
An average demand calculation procedure for calculating an average demand at each time when the demand is predicted,
A deviation calculating procedure for calculating a deviation at the current time based on a deviation between the average demand calculated by the average demand calculating procedure and the actual demand;
Predicted demand amount calculation for calculating the demand amount at the prediction target time based on the average demand amount calculated by the average demand amount calculation procedure with respect to the prediction target time and the deviation calculated with respect to the current time by the deviation calculation procedure A demand forecasting program characterized by causing a computer to execute the procedure.

Images