JP2015188280A - Apparatus controller and apparatus control method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To predict the power consumption of an apparatus with high precision.SOLUTION: The apparatus controller includes an in-unit-period total-amount-of-power-consumption prediction section that predicts the total amount of power consumption for a unit period on the basis of the total amount of power consumption of a plurality of apparatuses for the last unit period corresponding to a unit period which is a prediction target, and a day-of-week coefficient calculation section that obtains a day-of-week coefficient which is a ratio of a total amount of power consumption of the plurality of apparatuses on each day of week to that on all days of weeks.

Description

  The present invention relates to device control technology.

  There are methods of directly controlling the device so that the electricity charge (power consumption) in a certain period becomes a target value set in advance by the user, or indirectly controlling the device by notifying the user of advice.

  Patent Document 1 describes a device control method using a control line in order to achieve a target power reduction amount. It also describes using a predicted value when resetting a target value.

JP 2012-191717 A

  In Patent Literature 1, day-of-week dependency is not considered for the integrated power consumption for each time zone. It is an object of the present invention to predict the integrated power amount with higher accuracy.

  According to one aspect of the present invention, the total consumption within a unit period for predicting the total power consumption of a unit period based on the total power consumption of a plurality of devices in the unit period of the previous year corresponding to the unit period to be predicted. In a power amount prediction unit, a day of the week coefficient calculation unit that obtains a day of the week coefficient that is a ratio of the total power consumption of a plurality of devices on each day to the total power consumption on all days, and the total power consumption prediction unit in the unit period By dividing the obtained total power consumption of the unit period by the number of days included in the unit period and multiplying by the day of the week coefficient calculated by the coefficient calculation unit corresponding to the day of the prediction target day, 1 There is provided a device control apparatus (information processing apparatus) characterized by including a daily total power consumption prediction unit that predicts a total daily power consumption.

  According to another aspect of the present invention, based on the total power consumption of a plurality of devices in the unit period of the previous year corresponding to the unit period to be predicted, the total power consumption in a unit period is predicted. A power consumption prediction step; a day of the week coefficient calculation step for obtaining a day of the week coefficient which is a ratio of the total power consumption of a plurality of devices on each day to the total power consumption on all days; and the total power consumption prediction step in the unit period By dividing the total power consumption of the unit period obtained in step by the number of days included in the unit period and multiplying by the day of the week coefficient calculated in the coefficient calculation step corresponding to the day of the prediction target day, There is provided a device control method (information processing method) comprising a daily total power consumption prediction step for predicting a total power consumption per day.

  According to the present invention, it is possible to predict the accumulated power amount of a device with higher accuracy.

It is a functional block diagram which shows the example of 1 structure of the apparatus control apparatus used in the apparatus control technique by embodiment of this invention. It is a flowchart figure which shows an example of the flow of the power consumption amount prediction process for every month by a power consumption amount estimation part. It is a figure which shows the example of the actual value and average temperature which were read and read the past monthly power consumption. It is a figure which shows the example of calculation of the total power consumption predicted value for one month by initial prediction. It is a figure which shows an example of how to select the object month used for calculation of temperature dependence correction | amendment. It is a figure which shows the example of calculation for calculating expected temperature. It is a figure which shows the example of the algorithm of the temperature dependence correction | amendment depending on temperature, and is a figure which shows the content of selection of the parameter used for calculation exemplarily. The functional block diagram which shows the example of 1 structure of the 1st prediction part which performs the integrated power consumption prediction process of the general apparatus among the power consumption prediction parts which perform the power consumption prediction process by the 3rd form of this Embodiment. It is. It is a figure which shows the power consumption amount of a prediction object day, and the calculation object day used for a prediction with a calendar. It is a flowchart figure which shows the flow of a process in the information processing part by this Embodiment. It is a functional block diagram which shows the detail of a memory | storage part. It is a functional block diagram which shows the example of 1 structure of the 2nd estimation part which performs the power consumption amount prediction process by this Embodiment. It is a figure which shows the example of creation of the power consumption data with respect to an operating environment. It is a flowchart figure which shows the 1st example of the power consumption prediction process based on on-off determination. It is a flowchart figure which shows the 2nd example of the power consumption prediction process based on on-off determination. It is a figure which shows the example of the process result by the 1st example of the power consumption prediction process of an on-off determination. It is a figure which shows the example of the process result by the 2nd example of the power consumption prediction process of on-off determination. It is a figure which shows an example of the integrated power consumption prediction method for every day by the 1st Embodiment of this invention. It is a figure which shows an example of the predicted value of the total power consumption of 1st by the 3rd form of this Embodiment. It is a figure which shows an example of the temperature and power consumption table created by calculating | requiring the power consumption of the eco cute depending on the temperature by the 5th Embodiment of this Embodiment. It is a figure which shows the example of the measurement data by the 5th Embodiment. It is a figure which shows the example of transition of the integrated electric energy for every time in object date. It is a flowchart figure which shows the flow of the process regarding a peculiar day. It is a flowchart figure which shows the flow of the process which selects D1. It is a figure which shows the specific example regarding the 1st determination method and the 2nd determination method for selecting a day with total power consumption average. It is a figure which shows the mode of selection by the 1st determination method which selects D1 based on the data of FIG. It is a figure which shows the mode of selection by the 2nd determination method which selects D1 based on the data of FIG. It is a flowchart figure which shows the flow of a calculation (consideration of a day-of-week ratio) process of the estimated total power consumption amount for every day. It is a figure which shows the example of calculation of the total power consumption predicted value for every day of the prediction object month (May 2013) calculated by the 1st method or the 2nd method. It is a figure which shows the total power consumption average value for every day of the week in March to May, 2012, and the total power consumption amount ratio of each day of the week in one week. It is a figure which shows the total power consumption average value for every day of the week in March to April, 2012, and the total power consumption ratio of each day of the week in one week. Estimated value of cumulative power consumption change every 15 minutes (curve 1), and the integrated power consumption change every 15 minutes of 2013/5/22 obtained by correcting this with the total power consumption prediction value of the day It is a figure which shows a predicted value (curve 2).

  Factors affecting the daily power consumption of devices such as home appliances are as follows. Therefore, in order to obtain the daily power consumption, it is necessary to consider the following points.

1) There are ambient environment factors such as temperature, weather, humidity, etc. Basically, ambient environment factors represented by temperature 2) Other than ambient environment factors such as temperature, for example, due to changes in family composition, etc. 3) Lifestyle factors such as day-of-week dependency The device control technology according to the embodiment of the present invention will be described in detail below with reference to the drawings.

(First embodiment)
As will be described below, in the device control technique according to the first embodiment of the present invention, the accumulated power consumption per day is predicted according to the following steps.

  1) Based on the total power consumption of 1 month (target value or forecast value) calculated by the algorithm for predicting the total power consumption of 1 month (unit period) Calculate a predicted value of total power consumption.

2) Based on the actual values (power consumption, temperature, etc.) in the past T days,
For example, the integrated power amount transition (for example, every 15 minutes) (the ratio of the integrated power amount every 15 minutes in one day) is predicted.

  3) By correcting the ratio obtained in 2) with the total power consumption obtained in 1), an estimated integrated power consumption value for each time zone in a day is calculated.

For example, assume that the integrated power amount at 23:45 in 2) (the integrated power amount from 0:00 to 24:00) is 10,000 Wh, and the total power consumption per day for the prediction target day in 1) is 13000 Wh. The estimated integrated power consumption value every 15 minutes in one day is
Ratio of integrated electric energy every 15 minutes obtained in 2) × (13000 Wh / 10000 Wh)
Can be obtained as

  FIG. 11 is a diagram illustrating an example of a method for predicting an integrated power consumption for each day as described above. In step S51 shown in FIG. 11, the total power consumption for one month is predicted. Further, in step S52, weighting by day of the week (day of the week coefficient) is obtained. In step S54, based on the values obtained in step S51 and step S52, the total daily power consumption considering the day of the week is predicted. On the other hand, in step S53, the transition of the integrated power consumption is calculated every unit time of the day, for example, every 15 minutes. In step S55, the total daily power consumption obtained in step S53 (B) is matched with the total power consumption obtained in step S54 (A), so that the integrated power consumption every 15 minutes can be calculated. Correct the transition.

Steps S51, S52, S53, S54, and S55 include a function unit that includes a total power consumption prediction unit for one month, a day-of-week coefficient calculation unit, a daily (integrated) power consumption amount transition prediction unit, and a total for each day. It corresponds to a power consumption amount prediction unit and an integrated power amount transition correction unit.

  Then, in step S55, for example, a predicted value of the transition of the integrated electric energy every 15 minutes can be obtained by matching the daily integrated value obtained in step S53 with the value obtained in step S54. Based on these values, the device can be controlled in step S56.

  As described above, according to the present embodiment, it is possible to accurately obtain the transition of the daily power consumption per unit time. Therefore, since control of equipment can be performed with high accuracy, there is an advantage that it is easy to obtain a guideline for realizing energy saving, for example.

(Second Embodiment)
Next, the second embodiment of the present invention will be exemplarily described with respect to the total power consumption prediction algorithm for one month (step S54 in FIG. 11) in step 1) of the first embodiment.

  FIG. 1 is a functional block diagram showing a configuration example of a device control apparatus used in a device control technique according to an embodiment of the present invention.

As shown in FIG. 1, the device control apparatus 1 is a HEMS management in a system (for example, a home energy management system (HEMS)) that can collectively control a plurality of devices such as home appliances and house equipment. It consists of parts. For example, the device control apparatus 1 includes a power consumption amount prediction unit 11 that predicts a power consumption amount, a device control unit 15, a control unit 17 that controls the whole, a storage unit 21 that stores various data and programs, A display unit 23 that displays a user interface and the like, and an input unit 25 that performs user input in a state where the user interface is displayed and the like are included. The control unit 17 includes a display control unit 17a that performs display control for displaying a user interface and the like. The storage unit 21 also includes a power consumption data storage unit 21a that stores past power consumption data and the temperature of the day. And based on the target value input from the input part 25, the apparatus control part 15 is comprised so that various household appliances etc. can be controlled automatically.

  A part of the above configuration may be realized by a portable terminal such as a television device or a smartphone. For example, the display control unit 17a may include an output unit that outputs a display control instruction signal for displaying a user interface on a portable terminal such as a television apparatus or a smartphone, and the form of the device control apparatus 1 is limited. Not.

  In addition, the power consumption of each device is measured, for example, by measuring the power consumption of each device with a “plug-type power sensor” or a “CT sensor” attached to a distribution board, and the measurement result is represented by the HEMS management unit of the HEMS. You may make it provide to.

  Note that the components of the device control apparatus 1 in FIG. 1 do not have to be integrated. For example, the power consumption amount is predicted by a server connected to the device control apparatus 1 through a network and the like. You may make it provide the information. Such a system is also within the scope of the present invention.

  FIG. 2A is a flowchart illustrating an example of the flow of the power consumption amount prediction process for each month by the power consumption amount prediction unit 11. The power consumption amount prediction process is not limited to these examples, and can be selected from various prediction algorithms.

First, when the process is started (Start: Step S1), in Step S2,
The actual value of the power consumption data is read from the power consumption data storage unit 21a by the control unit 17. Also, environmental data such as average temperature is read out.

  FIG. 2B is a diagram illustrating an example of past actual power consumption values and average temperatures read in step S2.

  Hereinafter, for example, the state of initial prediction in the case where the total power consumption in October 2013 is predicted for the total power consumption for one month is shown. This algorithm obtains a predicted value on the assumption that the tendency when the total power consumption is monthly dependent is the same as the trend until last year.

  When estimating the total power consumption of a month, it can be estimated that it will basically be the same power consumption as the same month of the previous year, but the number of families and home appliances are increasing and decreasing compared to the previous year. there is a possibility. The initial prediction of the power consumption is taken into account the effect.

  In addition, because some home appliances have different usage due to temperature differences, the initial prediction is corrected to take into account the effects of temperature differences, and two steps are taken for one month. Estimate total household power consumption.

  The initial forecast is obtained by multiplying the previous month actual value by the increase / decrease ratio of the total power consumption in the same month of the previous year and the previous month of the previous year. This process is a time-dependent correction that does not depend on temperature, for example, the number of families or the type and number of household appliances used, and does not depend on temperature. It can also be said. Hereinafter, this is referred to as initial prediction.

FIG. 2C is a diagram illustrating a calculation example of a predicted value of total power consumption for one month based on initial prediction.
Here, the total power consumption for one month with the target year = y and the target month = m is represented by Ey, m [Wh]. The initial prediction E * Y, M [Wh] with respect to the total power consumption for one month in the prediction target month (Y month M month) is obtained by the following formula.
When the data of the previous month of the target month is available
E * Y, M = EY-1, M ・ (EY, M-1 / EY-1, M-1)
When the data for the previous month of the forecast target month is not available
E * Y, M = EY-1, M ・ (EY, M-2 / EY-1, M-2)
A specific example of initial prediction will be described below.

The initial prediction is, for example, correction using the following formula. In this example, the change rate (B / C) of the total power consumption, which is the actual value B in the same month the previous year (October 2012) and the actual value C in the previous month (September 2012) immediately before that, is the previous month (actual) The value obtained by multiplying the actual value D of the value (September 2013) is the predicted value for the next month.
A = D × (B / C)

  For example, in August 2013, D (August 2013) = 797989 × (395642/478378) = 659976. In September 2013, it will be 886250. The predicted value A in October 2013 is obtained as 565285 × (451215/531289) = 480936. This algorithm is not limited to the above as long as it is an algorithm that obtains a predicted value assuming that the tendency when the total power consumption value is monthly dependent is the same as the trend up to last year.

  Next, in step S4, temperature dependent correction depending on the temperature is performed. The temperature-dependent correction is a prediction value E that takes into account the amount of change in the total power consumption relative to the average temperature before and after the predicted temperature of the target month in the past data with respect to the predicted value A after the initial prediction. It is.

FIG. 2D is a diagram illustrating an example of how to select a target month to be used for calculating temperature-dependent correction.
A correction is performed for the predicted value obtained by the initial prediction, taking into consideration the effect of the difference in temperature.

In addition, the average temperature for one month with the target year = y and the target month = m is expressed as Ty, m [° C.]. In addition, it is assumed that the predicted average temperature for one month of the target month (Y month M month) is given by T * Y, M [° C.].

  Since the previous year's data includes data on average temperature and average power consumption, the trend of average temperature and average power consumption can be grasped by referring to the data. However, since the relationship between the average temperature and the average power consumption is not a linear relationship, the month whose average temperature is close to the expected average temperature of the target month for which the power consumption is predicted is extracted from past data. Then, the power consumption amount of the prediction target month is corrected by referring to the information on the power consumption amount of those months.

More specifically, as shown in FIG. 2D, among the average temperatures T _{Y−1, m} (M−2 ≦ m ≦ M + 1) of each month during the two months before and after the same month of the prediction target month, T _{Y-1, M} and _TY-1,
The month m in which T * _{Y, M} is sandwiched between _m and | T _{Y−1, M} −T _{Y−1, m} | is minimized is selected. (The month is hereinafter referred to as m.)
However, the priority for selecting m is as follows.
1) Select m in the range of M ± 1.
2) Select M within the range of m ± 2.

The predicted value E ** _{Y, M,} which is obtained by performing temperature-dependent correction on the initial predicted E * _{Y, M} for the total power consumption for one month of the target month (Y year M month), is obtained by the following formula.
E ** _{Y, M} = E * _{Y, M}・ α
α = {E _{Y-1, M} + (E _{Y-1, M} −E _{Y-1, m} ) ・ (T * _{Y, M} −T _{Y-1, M} ) / (T _{Y-1, M} −T _{Y-1, m} )} / E _{Y-1, M}
However, if M satisfying the conditions is not found when selecting 1) and 2) above, correction based on temperature is not performed and E ** _{Y, M} = E * _{Y, M.}

A specific example of temperature dependence prediction will be described below.
FIG. 2E is a diagram illustrating a calculation example for calculating the predicted temperature. As shown in FIG. 2E, the average temperature of each month is calculated from the weather data for the past N years in the prediction target area. In the example shown in FIG. 2E, the average temperature for each month calculated from the data for the past five years is averaged over five years as the predicted temperature of the prediction target month. For example, when the prediction target month is May 2013, the predicted temperature of the prediction target month is (22 + 23 + 19 + 25 + 24) /5=22.6° C. in the example of FIG. 2E. Furthermore, the predicted temperature coefficient Tk of this year may be added to this value. As the coefficient Tk, for example, a difference from the annual temperature of each month provided by the Japan Meteorological Agency or a weather information company is used. If the month is 0.5 ° C higher than the average year, Tk = 0.5 should be added to the predicted temperature to make 23.1 ° C the expected temperature. However, this calculation method is exemplary, and is not limited to this method. For example, prediction data such as a predicted temperature provided by the Japan Meteorological Agency or a weather information company may be used.

  FIG. 2F is a diagram illustrating an example of an algorithm for temperature-dependent correction that depends on the temperature, and is a diagram exemplarily showing how parameters used for calculation are selected.

  In this process, when there is data for the previous year, the correlation between the average temperature and the total power consumption for each month of the previous year is used. The total power consumption of the prediction target month is predicted by correcting the actual value of the same month of the previous year using the predicted temperature of the prediction target month.

In FIG. 2F, in order from the left, the first column is the same month of the previous year in October 2013. 201
2 years and October are shown, and the second column shows T1, that is, the average temperature (from FIG. 2B) of the same month of the previous year. The third column shows the power consumption value (actual value October 2012) E1 of the same month of the previous year. The fourth column shows the months before and after the same month in the previous year used for temperature dependence correction. The fifth column shows the average outside temperature T2 in the months before and after the same month of the previous year. The sixth column shows the actual value E2 of power consumption in the months before and after the same month of the previous year. The seventh column is the predicted temperature T0 of the prediction target month, which can be obtained from the past data in FIG. 2B, for example, forecast data from the Japan Meteorological Agency may be used. Here, a value of T0 = 15 ° C. is used.

The eighth column is an estimated value E0 of power consumption in the month before and after the same month of the previous year when the temperature T0 = 15 ° C., and can be obtained using the following equation (1).
E0 = {(E1-E2) / (T1-T2)} × (T0-T1) + E1 (1)

The ninth column is the absolute value of the difference between T1 and T2. The tenth column is the value of E0 / E1. Then, the temperature dependence correction is performed according to the following procedure.
1) In (T1, E1) October and (T2, E2) September, see if T0 is between T1 and T2, and if | T1-T2 |
2) In (T1, E1) October and (T2, E2) November, see if T0 is between T1 and T2, and if so, calculate | T1-T2 |.
3) In step 1) and 2), formula (1) is calculated with the combination of months with smaller | T1-T2 |. By interpolating with a smaller value of | T1-T2 |, a better interval can be obtained. If both steps 1) and 2) are NG (T0 is not included between T1 and T2), proceed to the next step.
4) In (T1, E1) October and (T2, E2) August, see if T0 is in T1 and T2, and if there is, calculate | T1-T2 |.
5) In (T1, E1) October and (T2, E2) December, see if T0 is in T1 and T2, and if there is, calculate | T1-T2 |.
6) Equation (2) is calculated for the combination of months with the smaller | T1-T2 | in steps 4) and 5). If both steps 4) and 5) are NG, the temperature dependent correction value = A × 1.
E (secondary correction value) = A × [{(E1-E2) / (T1-T2)} × (expected temperature T0-T1) + E1] / E1] = A × E0 / E1
(2)

  Thus, the predicted temperature T0 of the target month is obtained, T0, the average outside temperature T1 of the same month of the previous year, the actual power consumption value E1, and the average outside temperature T2 of each month before and after the same month of the previous year, the actual power consumption. Based on the value E2, an estimated value E0 of power consumption in the months before and after the same month in the previous year in consideration of the temperature when the predicted temperature T0 of the target month is calculated is obtained for each month, and E0 / E1 is set as the initial predicted value A By multiplying, the temperature dependent correction value, that is, the total power consumption E for one month in consideration of the temperature dependency can be obtained with high accuracy.

(Process when total power consumption of this month is not found)
In the above example, as a method for predicting the total power consumption in October 2013, for example, a case where all the data of September 2013, which is the previous month of the same year, has been described. When predicting the total power consumption for next month, the data for September 2013, which is the month before the same year, are often not available.

  In such a case, in the initial prediction, the increase / decrease ratio (B / C) of the total power consumption, which is the actual value B in the same month of the previous year (October 2012) and the actual value C ′ of the month before the previous year (August 2012). A value obtained by multiplying ') by the actual value D' two months before the same year (actual value August 2013) can be used as a predicted value for the next month. Further, in the temperature dependent correction, the correction may be performed by step 4) -6) without performing the above step 1) -3).

  In the process for calculating the predicted power consumption, a simpler process or a more complicated process may be used. However, the process described above enables accurate prediction.

  Next, in step S5, the total power consumption of the target month calculated in step S4 is stored in the power consumption data storage unit 21a, and the process ends (step S6: end).

  As described above, according to the present embodiment, it is possible to accurately obtain the total monthly power consumption in the future based on data such as past power consumption.

  Next, based on the above result, a prediction algorithm for the accumulated power consumption transition in one day will be described.

  Prediction of the integrated power consumption transition every 15 minutes in one day is performed by the following method. Based on the algorithm for predicting peak power consumption, it was applied to the algorithm for predicting integrated power.

1) The following value is obtained for the fitting date which is the date one week before the prediction target date.
Note that t is time and the unit is 15 minutes. L is the last 28 days in the period.
A1 (t): Average value of accumulated energy from 0:00 to t minutes on each day of the period (L) before the fitting target date
B1 (t): Average value of accumulated energy from 0:00 to t minutes on the same day of the week as the forecast date (L) before the fitting date
C1 (t): Average value of accumulated power from 0:00 to t minutes after the last 7 days before the fitting date
D1 (t): Total power consumption every t minutes from 0:00 on the fitting target day 1) Using A1 (t), B1 (t), C1 (t), D1 (t) calculated in 1)
e (t) = D1 (t) − (a ・ A1 (t) + b ・ B1 (t) + c ・ C1 (t)) (a + b + c = 1, a ≧ 0, b ≧ 0, c ≧ 0)
Sum of squared errors in one day Σe ² (t)
Find a, b, and c that minimizes.

2) The following value is calculated for the prediction date.
A2 (t): Average value of integrated power consumption every t minutes from 0:00 in the period (L) before the forecast date
B2 (t): Average value of accumulated energy every t minutes from 0:00 on the same day of the week as the forecast date (L) before the forecast date
C2 (t): Average value of integrated power consumption every t minutes from 0:00 on the last 7 days before the forecast date 3) Integrated power every t minutes from 0:00 on the forecast date using the following formula Determine the quantity D2 (t).
D2 (t) = a ・ A2 (t) + b ・ B2 (t) + c ・ C2 (t)
The target of prediction is the total power consumption (integrated value) of the entire house. The total power consumption of the entire house here is the sum of all the child breakers (each room breaker, dedicated breaker) to which the devices that consume power are connected. The predicted value obtained is the accumulated power amount from 0:00 to time t every 15 minutes. The output result is as shown in FIG.

(Third embodiment)
A third embodiment of the present invention will be described. As shown in step S53 of FIG. 11, the accumulated power amount transition (1 day) for each time zone (for example, every 15 minutes) of the prediction target day from the actual value (power consumption, temperature, etc.) in the past T days in step 2). The ratio of the integrated electric energy every 15 minutes) is predicted.

In the present embodiment, the power consumption of all devices is obtained by the method described below.
FIG. 12 is a diagram illustrating an example of prediction data of changes in the daily power consumption. As shown in FIG. 12, the predicted value of the total power consumption per day is indicated by the value of the integrated predicted power output for the date and time, for example, every 15 minutes with respect to the date and time.

  FIG. 3 is a functional block diagram illustrating a configuration example of the first power consumption prediction unit 11a that performs the integrated power consumption prediction process of the device in the power consumption prediction unit 11 that performs the power consumption prediction process according to the present embodiment. is there. FIG. 4 is a diagram showing a prediction target date of the power consumption value and a calculation target date used for prediction in a calendar.

  The first power consumption amount prediction unit 11a includes a power consumption amount reading unit 11a-1 that reads power consumption values, a statistical processing unit 11a-2 that performs statistical processing related to information processing of power consumption values, and different power consumption values. Based on the statistical value for which the dependency is obtained, the weighting coefficient for the dependency is obtained, that is, the coefficient calculation unit 11a-3 for obtaining the weighting coefficient for the dependency of the period or the different period, and statistical processing and coefficient calculation are performed. As a result, the power consumption predicting unit 11a-4 that predicts the power consumption value based on the power consumption value is included.

  The prediction target day of the power consumption value is, for example, July 28 (Saturday), and the power consumption value measured in the previous period is used for the prediction. The extent of going back to the past can be arbitrarily selected, for example, one month ago. The prediction reference date is, for example, July 20 immediately preceding the same day of the week as the power consumption value prediction target date.

FIG. 5 is a flowchart showing a process flow in the power consumption prediction unit 11a according to the present embodiment.
The processing is started (Start: Step S11). From the storage unit 21 that stores the daily power consumption value together with the date and day of the week, the power consumption amount reading unit 11a-1 is the same day of week as the prediction target day. As a reference date, the first power consumption value of the past on the same day before the reference date and the second of the most recent x days (x is a positive value) before the prediction target date before the reference date. The power consumption value and the third power consumption value on the y (y> x) day longer than the x day before the prediction target date are read (step S12). Next, the statistical processing unit 11a-2, the first statistical value that statistically processed the first power consumption value, the second statistical value that statistically processed the second power consumption value, and the third A third statistical value obtained by statistically processing the power consumption value is obtained (step S13). Next, in step S14, the coefficient calculation unit 11a-3 obtains a weighting coefficient for the first statistical value, the second statistical value, and the third statistical value. In step S15, the power consumption prediction unit 11a-4 calculates the power consumption on the prediction target day based on the weighting coefficient (end in step S16).

Constants a, b, and c for obtaining the prediction result are obtained. (In the following calendar example, the coefficient for obtaining the de-measurement results up to the 20th can be obtained as follows (for example, the calendar example in FIG. 4 uses the data up to the 20th and compares it with the 21st). ).

A (t): Average power consumption value in the second period in the past (for example, from July 1st to July 20th of the reference date)
B (t): Average power consumption value in the past designated day of the week (for example, July 7 and July 14)
C (t): Average power consumption value in the past first period (a certain period, for example, from July 14 to the reference day 20)
D (t): Power consumption value on the target day (the day one week before the forecast date) for fitting the forecast results
D (t) = aA (t) + bB (t) + cC (t)
Calculate the combination of a, b, and c that is closest. Here, t is time.

The power consumption value is calculated from the constants a, b, c and A (t), B (t), and C (t) for the prediction target date. In FIG. 4, data up to 27th is used and 28th is predicted.

  Note that, for the above-described period and designated day of the week, general statistical processing such as excluding days with prominent values or changing the period depending on the obtained value can be applied.

Further, instead of obtaining the power consumption value as shown in FIG. 5, the time dependence of the power consumption value can be obtained. In this case, the statistical value and the coefficient value may be obtained for each time zone or every hour, or after dividing into on time and off time using threshold power consumption, etc., the on period and off time Statistical processing may be performed during the period.
Further, an average value may be used as the statistical value, and an extrapolated value may be used as the statistical value instead.

  As described above, according to the present embodiment, the weighting between the relatively recent value and the statistical value having a longer value among the day-of-week dependencies is obtained, and this weighting is calculated based on the consumption of the prediction target day. By using it when obtaining the power value, it is possible to perform prediction in consideration of day-of-week dependency and its period dependency.

  For example, when January 1, 2013 is the target day, D (t) is as shown in FIG. The total power consumption of the entire house is defined as the total power consumption of the target device. These values are used as the integrated power amount from 0:00 to time t.

In the above prediction method, a, b, and c are calculated by a sufficient method extending the dichotomy.
It is also possible to calculate a, b, and c by multiple regression instead of a sufficient method.
As described above, the power consumption of all devices can be obtained.

FIG. 16 is a flowchart showing the flow of processing relating to a peculiar day. This process detects a day when the amount of electricity used is abnormally high or abnormally low, and excludes the detected day from the prediction base. In each of the above processes, the fitting target date D1 is a day corresponding to one week before the prediction target date. However, for example, when D1 falls on a singular day, there is a problem that a prediction error becomes large.

Therefore, in order to solve this problem, the selection process of D1 is optimized.
As shown in FIG. 16, the process is started in step S61, and the target date of D1 is selected in step S62. Next, in step S63, coefficients a, b, and c are calculated based on the selected D1. In this case, singular days are excluded. Then, in step S64, based on a, b, and c, the predicted value is calculated by excluding the singular day, thereby ending the process (step S65).
In step S62 described above, D1 is selected by the process shown in FIG.

  As shown in FIG. 17, the process starts in step S71, and in step S72, candidate dates for D1 in the past T days are selected. Here, the candidate date of D1 is the same day one week before the prediction target date in the period T, the same day two weeks ago,..., The same day n weeks ago.

Next, in step S73, D1 is determined by the following processing.
The example of the 1st determination method for D1 is the method of selecting the day with the total power consumption average among each day which is the same day in the past (for example, 4 weeks) as a prediction target day as D1. . Alternatively, the median value may be used.

In the example of the second determination method for D1, D1 is in principle a day corresponding to one week before D2. However, if it is not included in the average power consumption ± X% (here X = 20) of the same day in the past (for example, 4 weeks) from the forecast target day, the day is the same day one week before If this day is also not included in the average power consumption ± X%, a method of selecting a day that is the same day of the week before one week may be used.

  FIG. 18 is a diagram illustrating a specific example related to the first determination method and the second determination method, taking as an example the case where the prediction target date D2 is November 22, 2012. Here, for example, the D1 candidate date is a date that is (7 × N) days back from D2, and is November 15, 2012, November 8, 2012, or November 1, 2012. FIG. 19 is a diagram showing a state of selection by the first determination method for selecting D1 based on the data of FIG. It is a figure which shows the difference of the value in D1 candidate day, the total power consumption for every candidate day, and D1 candidate day average value and D1. In FIG. 19, it can be seen that November 8, 2012 having the smallest absolute value of the difference is preferably the D1 candidate day.

  FIG. 20 is a diagram showing a selection state by the second determination method for selecting D1 based on the data of FIG. It is a figure which shows the ratio of the difference of the value in D1 with respect to D1 candidate day, the total power consumption for every candidate day, and D1 candidate day average value. Similarly to the case of FIG. 19, it can be seen that in FIG. 20, it is preferable to set November 8, 2012 having the smallest difference ratio as the D1 candidate date. By such a method, preferable D1 can be determined excluding the exclusion date of D1.

  In the third embodiment, the power consumption is predicted without distinguishing between the types of devices. However, in the following, the power consumption is predicted by a more accurate method for the air conditioner and eco-cute, respectively. An example of processing for predicting the device is shown by the method according to the third embodiment.

And the predicted value of the transition of the power consumption per day is calculated | required as follows, for example.
Predicted value of daily power consumption transition = predicted power consumption of air conditioner + predicted power consumption of Ecocute + predicted power consumption of other devices

  In the fourth embodiment, an example will be described in which the air conditioner is predicted, and in the fifth embodiment, the ecocute is predicted by a method different from that of the third embodiment. When the predicted value is obtained for only one of the methods by a method different from that of the third embodiment, the above equation may be applied only to the one obtained by a different method.

  Ecocute is a registered trademark that uses carbon dioxide instead of CFC as a refrigerant in an electric water heater that can boil hot water with the heat of air using heat pump technology.

(Fourth embodiment)
A fourth embodiment of the present invention will be described.
First, regarding the air conditioner apparatus, a process for predicting the integrated power consumption without using the method in the third embodiment will be described.

  As shown in FIG. 6A, the storage unit 21 that stores the measurement value stores the power consumption value measured by the power consumption measuring instrument or the like or the time change thereof together with the date, day of the week, etc. in units of 24 hours, for example. Power consumption amount storage unit 21-1, an operation state data storage unit 21-2 that stores the operation state of the air conditioner, an on / off period storage unit 21-3 that stores an on period and an off period of the air conditioner, and an air conditioner on state An on / off power storage unit 21-4 that stores the power of the period and the power of the off period, and a time change function (characteristic) storage unit 21-5 that stores a time change (time change (characteristic) function) of the power consumption of the air conditioner. And have.

Details of the power consumption prediction process will be described in detail below.
FIG. 6B is a functional block diagram illustrating a configuration example of the second power consumption amount prediction unit 11b that performs the power consumption prediction process according to the present embodiment. The calendar is used to refer to the power consumption prediction date and the calculation date used for prediction shown in FIG.

  The second prediction unit 11b illustrated in FIG. 6B reads the on period, the off period, and the like of the air conditioner device (apparatus) from the driving state data storage unit 21-2, the on / off period storage unit 21-3, and the like of the measurement value storage unit 21. A statistical processing unit 11b-2 that performs a statistical process related to time-dependent information processing on and off, and obtains an on probability of the device in a certain period, an on probability obtained by performing the statistical process, An on / off period determination unit 11b-3 that determines an on / off period of the air conditioner based on the threshold setting unit 11b-3-1 and an on / off power storage unit 21-4 or a time for each of the on period and the off period. The on / off power allocation unit 11b-4-1 that allocates the power consumption stored in the change function storage unit 21-5 and the on / off power allocation, etc. It has a power consumption prediction portion 11b-4 to make predictions of the power, the.

  The statistical processing unit 11b-2 obtains a statistical value (for example, average value: on probability) obtained by statistically processing the on period and the off period of the device in a certain past period. As shown in FIG. 4, for example, an ON probability is obtained based on an ON / OFF period in a certain period of one month (for example, July), a certain week, and a certain day. The on-probability (off probability) can be obtained from the operation status data storage unit 21-2 or the on-off period storage unit 21-3 of the measurement value storage unit 21a, for example, the on-period and off-period of the device.

  The prediction target day for predicting power consumption is, for example, July 28 (Saturday), and the ON probability measured in the previous period is used for the prediction. The extent of going back to the past can be arbitrarily selected, for example, up to one month ago. Taking an average value of on / off, for example, if on is 3 days and off is 4 days, the on probability is 3/7.

  FIG. 8A is a flowchart showing a first example of a power prediction method according to the present embodiment. FIG. 9 is a diagram illustrating an example of a processing result according to the first example.

  Here, for example, the on / off driving status data in which the reading unit 11b-1 specifies the on / off period of the air conditioner depending on the time from the driving status data storage unit 21-2, the on / off period storage unit 21-3, and the like. Is read. Then, time dependency such as the ON probability of a certain period is obtained by statistical processing as necessary (see FIG. 9A).

  As shown in FIG. 8A, when the information processing is started from the state in which the ON probability data in FIG. 9A is obtained (step S21), in step S22, the on / off period determination unit 11b-3 turns on / off the device. The period is determined (see (b) of FIG. 9A). Next, in step S23, the on / off power allocation unit 11b-4-1 applies the on-time power consumption during the on-time and the off-time power consumption during the off time with reference to the value of the on / off power storage unit 21-4. By doing (see (c) of FIG. 9A), in step S24, the power consumption amount prediction unit 11b-4 predicts the power consumption of a certain period, for example, a prediction target day (such as July 27 in FIG. 3). .

As described above, by predicting or determining the time distribution between the on-period and the off-period of the device and determining the power consumption based on this, it is possible to predict the power consumption with high accuracy with a simple process. The on-time power consumption and the off-time power consumption may be created based on actual measured values of the devices, or may be created based on design values such as rated values. The power consumption may be obtained by assigning the on power only to the on period and setting the off period to zero. The on / off determination unit 21-3 may determine only the on period (or only the off period). In this case, the other period is an off period (or on period). The process ends (step S25).

  Here, the threshold value in the on / off determination is set to 0.5, but the threshold value may be adjusted according to the season, latitude and longitude. In a warm season or a position close to the equator, it is possible to adjust the threshold value to a small value so that the on period of cooling is extended. Further, it may be changed depending on the type of device.

  FIG. 10 is a diagram illustrating a second example of the power consumption prediction process according to the present embodiment. FIG. 8B is a flowchart showing the flow of processing according to the second example, and corresponds to FIG. 8A. In this second example, the power consumption at the time of on and the power consumption at the time of off are respectively calculated based on the measured values of the device, and the power consumption information for each use condition (operating environment) of the device is taken into account. The obtained value is used. Here, as an example, an example will be described in which the power consumption value for each use condition of a different device depends on temperature prediction data or actually measured values. FIG. 7 is a diagram illustrating an example of creating power consumption data for the operating environment.

  As the power consumption data for the operating environment, for example, as shown in FIG. 7, a power consumption database (lookup table) for the operating environment is created based on the past power consumption data. For example, in the case of an air conditioner, data when it is determined to be on is extracted. For example, an average value for every 15 minutes of “time after activation” after activation of the air conditioner is obtained for “temperature” and “power consumption” as shown in FIG. Here, the average value of the power consumption for 15 minutes from 0:00 to 0:14 after the start-up time is statistically obtained such that the average temperature is 24.9 ° C. and the average power consumption is 1000 W. Here, in order to obtain a power consumption curve for “time after start (15 minutes)” and “temperature”, data in the range of “time after start (15 minutes)” and “temperature” ± 1 ° C. are picked up. It is better to eliminate the influence of sudden data by obtaining the average.

  When there is a missing data, linear interpolation and extrapolated values may be interpolated for the temperature.

  Then, as shown in FIG. 7B, for example, a lookup table is created in units of 1 ° C., for example, as indicated by reference sign P, the elapsed time (time after startup) is 0:00. The power consumption is 1000W.

This is the average power consumption from the time 0:00 to 0:14 after the start of the temperature when the temperature for 15 minutes from the time 0:00 to 0:14 is 24 ° C to 26 ° C It has become. As a result, when it is predicted that the temperature is 26 ° C. and 15 minutes have elapsed since the start at 7:00 from the start of the prediction target day, the power can be predicted and output to be 500 W. You can create a database such as a lookup table. Although the data format and the like are not limited to the above-described forms, power consumption data based on the operating environment may be created in this way.

FIG. 10 is a diagram illustrating an example of a processing result according to the second example of the power consumption prediction process.
As shown in FIG. 8B, the process is started (step S31), and the past on period and off period of the device are read in step S32. Then, as shown in step S33, the time dependence of the on probability is divided into on / off periods for each day by the on / off prediction method as described above (FIG. 10 (a)).

  Next, as shown in FIG. 10B, the power consumption is predicted by applying the power consumption during the on period and the power consumption during the off period during the off period. When the power consumption changes depending on the temperature, for example, as shown in FIG. 10 (c), in the air conditioner, the power consumption is 300 W when the temperature is 25 ° C., and the power consumption when the temperature is 30 ° C. Is 600 W, based on the change in temperature depending on the time, as shown in FIG. 10 (d), the power consumption of 300 W is consumed in the on period of 25 ° C. and the power consumption of 600 W is consumed in the on period of 30 ° C. Electric power is applied to predict power consumption (step S34), and the process ends (step S35).

  In this way, it is possible to improve the accuracy of power consumption prediction based on power consumption information for each usage environment (condition) of the device such as temperature.

  The predicted power consumption value of a device such as an air conditioner every 15 minutes predicted as described above is converted into a power consumption amount. Based on the converted power consumption predicted value as shown in FIG. 12, for example, a value integrated from 0:00 to time t is output as the integrated predicted power amount.

(Fifth embodiment)
A fifth embodiment of the present invention will be described.
Next, regarding eco-cute, a process for predicting integrated power consumption without using the method according to the third embodiment will be described. Basically, for eco-cute, as with an air conditioner, the integrated power consumption is calculated based on the on / off probability.

  For example, FIG. 13 is a diagram illustrating an example of an air temperature / power consumption table created by obtaining eco-cute power consumption depending on air temperature. The method of extracting actual measurement data input to the temperature / power consumption table shown in FIG. 13 is as follows.

First, the outside air temperature (° C.) when the on flag is set (when the power data exceeds the threshold value) is recorded.
Next, the average power consumption (W) between “time after activation (time after turn-on)” is recorded.

  FIG. 14 is a diagram illustrating an example of actually measured data. For example, it is possible to measure the elapsed time after the activation of Ecocute every minute, the temperature at the elapsed time, and the power consumption value. This value is stored in the storage unit 21.

  The value stored as the temperature / power consumption table is, for example, the temperature is rounded off to 20.5 ° C to 21 ° C, and the power consumption is the time from when Ecocute is turned on until it is turned off (0: 00 to x: xx) The average power consumption is calculated to be 1100W.

  In the temperature / power consumption table of FIG. 14, for example, the average of the data of 19.5 ° C. to 20.4 ° C. that can be approximated to 20 ° C. and the power consumption of 0:00 at 0:00 and the outside air temperature at 20 ° C. Assign a value.

  Each time this table is updated, it takes time to call all the data from the database storing the raw data, and the processing load increases. Therefore, it is preferable to store the “average value” and the “number of data for which the average value is obtained” for each of the stored cells in FIG.

  For example, if the number of data is N and the average value is A and the new data is B, the average value is updated to (N × A + B) / (N + 1). Thereby, data can be updated at any time with a simple process.

If there is a missing data in the table created by extracting the measured data, it is preferable to compensate for the temperature by linear interpolation processing and extrapolation processing.

  Next, for example, the predicted power consumption can be obtained from the predicted on / off data and the predicted temperature by the following method.

Then, the predicted power consumption obtained as described above is converted into the predicted power consumption as follows in the same manner as in the case of an air conditioner.
15 minutes of predicted power consumption (Wh) = 15 minutes of predicted power consumption (W) / 60 × 15
The integrated predicted power consumption is an integrated value of the predicted power consumption from 0:00 to time t.

  As described above, the prediction method applicable to all the devices described in the third embodiment, the method for performing on / off prediction for the air conditioner as described in the fourth embodiment, or the fifth embodiment. As described above, it is possible to perform more accurate power consumption prediction in combination with a method of performing on / off prediction for eco-cute. For example, the integrated predicted power amount can be obtained as an integrated value of the predicted power consumption amount from 0:00 to time t.

(Sixth embodiment)
A sixth embodiment of the present invention will be described.
Next, in the sixth embodiment, as shown in step S54 based on the total power consumption predicted value for each month obtained in step S51 of FIG. 11 and the day of the week coefficient calculated in step S52 as described above. A process for predicting the total power consumption per day (considering the day of the week coefficient) will be described.

FIG. 21 is a flowchart showing the flow of the calculation of the total power consumption prediction value for each day (considering the day of the week ratio) in step S54.

  First, in step S101, the process is started (Start), and in step S102, it is determined whether or not the previous year's data is stored in the power consumption data storage unit 21a. If it is determined that the previous year's data is stored (Yes), in step S103, the total power consumption of each day of the week in the same month in the previous year and the month before and after (the first method) or the previous two months (the second method) The amount average value is obtained.

  On the other hand, if NO is determined in step S102, the process proceeds to step S104, and it is determined whether or not data from one month to one year before the month to which the prediction target day belongs is stored. In the case of Yes in step S104, an average value of total power consumption for each day of the week before and after that (first method) is obtained in step S105.

  Next, the process proceeds from step S103 or step S105 to step S106, and the total power consumption for each day is obtained in consideration of the ratio of the total power consumption average value for each day to the total power consumption average value for all days.

  On the other hand, if NO is determined in step S104, the process proceeds to step S107, and the process ends without setting the target value.

  FIG. 22 is a diagram illustrating a calculation example of a predicted total power consumption amount for each day of a prediction target month (May 2013) calculated by the first method or the second method. The predicted total power consumption for each day indicates the predicted value for each month, the day of the week ID indicating the day of the week, and the predicted power consumption for each day determined by the first method and the second method, respectively. FIG.

FIG. 23 is a diagram showing an average value of total power consumption for each day of the week from March to May 2012 and a ratio of total power consumption for each day of the week in one week.

  FIG. 24 is a diagram showing an average value of total power consumption for each day of the week in March to April 2013 and a ratio of total power consumption for each day of the week in one week.

  The day ID is a flag indicating the day of the week. For example, 1 indicates Sunday, 2 indicates Monday,..., 7 indicates Saturday.

  In FIG. 22, the predicted value for the total power consumption for one month uses the value after the temperature-dependent correction of the above-mentioned method for predicting the power consumption for each month.

In FIG. 23, “total power consumption_average value” indicates the total power consumption average value for each day of the week from March to May 2012.

The day-of-week coefficient indicates the power amount ratio of each day of the week in the prediction target month. In FIG. 23, for example, for day ID 1 (Sunday), it can be obtained by the following formula.
(Total power amount of day ID 1_Average value) / (Total of all power amount average values for all days)
= 8905 / (8905 + 97977.4 + 1003.7 +... +10107.3) = 0.13.
The day of the week coefficient in FIG. 24 can be obtained similarly.

In FIG. 22, the calculation method of the total power consumption per day of the prediction target month by the first method and the second method is as follows.
Predicted total power consumption per day = predicted total power consumption for the month to which the forecast target day belongs * (day of the week corresponding to the forecast target day / sum of the day of the week belonging to the forecast target month)
Therefore, the predicted value A calculated by the first method is A = B × (C4 / ((C4 + C5 +
C6) × 5 + (C7 + C1 + C2 + C3) × 4))
It is.

On the other hand, the predicted value D calculated by the second method is as follows.
D = B × (E4 / ((E4 + E5 + E6) × 5 + (E7 + E1 + E2 + E3) × 4)
It is.

  When the target value for one month does not exist, the prediction is not performed and, for example, a message (screen) for prompting the target value setting is displayed on the user interface.

  Next, by matching the total power consumption amount obtained in step S55 of FIG. 11, that is, the total power consumption amount obtained in step S53 (B), with the total power consumption amount obtained in step S54 (A), The transition of the integrated electric energy every 15 minutes is corrected.

FIG. 25 shows a predicted value (curve 1) of the integrated power consumption transition every 15 minutes and an integrated consumption every 15 minutes of 2013/5/22 obtained by correcting the predicted value with the predicted total power consumption of the day. It is a figure which shows the predicted value (curve 2) of electric energy transition. That is, it is obtained by the following formula.
Daily integrated power consumption transition = power transition obtained by processing in step S53 × (total power consumption obtained in step S54 / total power consumption obtained in step S53)

In the following steps, the transition of the integrated power consumption is predicted every 15 minutes.
1) Calculate the total power consumption per day for the total power consumption (target value or predicted value) for one month, taking into account the day of the week dependency.
2) Transition of integrated power consumption for each time zone (for example, every 15 minutes) from the actual values (power consumption, temperature, etc.) in the past T days (ratio of integrated power consumption every 15 minutes in a day, ( 0: 0
0-0: 15, 0: 00-0: 30,..., 0: 00-24: 00) is predicted.
3) By correcting the ratio obtained in 2) with the total power consumption obtained in 1), an estimated integrated power consumption value for each time zone in a day is calculated.

For example, assume that the integrated power amount at 23:45 in 2) (the integrated power amount from 0:00 to 24:00) is 10,000 Wh, and the total power consumption per day for the prediction target day in 1) is 13000 Wh. The estimated integrated power amount every 15 minutes in one day is obtained as the ratio of the integrated power amount every 15 minutes obtained in 2)) × (13000 Wh / 10000 Wh).

  As a more detailed example, for example, a prediction example of the transition of the integrated power consumption every 15 minutes of 2013/5/22 is shown.

  By the method described in step S53 of FIG. 11, the integrated power consumption amount transition is obtained every 15 minutes of 2013/5/22. This transition is shown in FIG.

  On the other hand, the total daily power consumption of 2013/5/22 obtained in step S54 (method 2) is 15267.2 Wh from FIG.

Here, the total power consumption in the integrated power consumption transition every 15 minutes obtained by the method of step S53 (for example, 11600 Wh as an example of an actual calculated value calculated by the sufficient method)
Is corrected to the total daily power consumption (15267.2 Wh) obtained by the method of step S54 (see arrow).

  And finally, the predicted value of the integrated power consumption transition every 15 minutes can be obtained as the predicted value of the power consumption of 2013/5/22. That is, the final value is matched with the daily total power consumption predicted value obtained in step S54, and the state of the transition is the shape of the transition obtained in step S53. , The coefficient (final value in step S54 / final value in step S53).

  As described above, by calculating the integrated power consumption prediction value for each unit time, using the control line based on this, the power consumption of the entire house is set to a preset target value. Control or advice notification can be made.

(Seventh embodiment)
A seventh embodiment of the present invention will be described. In this embodiment, in the sixth embodiment, the total power consumption for one month is first predicted, and when this is distributed as the predicted total power consumption for each day, every day passes. The actual value is accumulated. In the present embodiment, a total power consumption prediction correction function for correcting the predicted value is provided based on the actual value. First, the total power consumption on the prediction target day is calculated by the following formula.
Predicted total power consumption for the forecast target day = (predicted total power consumption for the month to which the forecast target day belongs −
(Actual value until the forecast target day) × (Electricity ratio of the day of the week corresponding to the forecast target day / Sum of the power ratio for all days of the week from the forecast target day to the end of the day belonging to the forecast target month)

  The total power consumption prediction value for the prediction target month is A [Wh] by the process of step S51 of FIG. 11, and the total power consumption prediction value for each day of the prediction target month is a1 by the process of step S54 of FIG. , A2,... An [Wh], respectively.

If the actual value on the first day of the prediction target month is r1 [Wh], (A-r1) [Wh] is redistributed in the remaining period (from day 2 to day n) according to the process of step S54. To do.
The redistributed predicted values are a2 ′, a3 ′,..., An ′ [Wh].

For example, if the predicted value on May 1, 2013 is X1 [Wh], X1 is obtained from the tables of FIGS.
X1 = (B-0) × (E5 / ((E4 + E5 + E6) × 5 + (E7 + E1 + E2 + E3) × 4)).

Next, assuming that the predicted value on May 2, 2013 is X2 [Wh] and the actual value on May 1, 2013 is Y1 [Wh], from the tables of FIG. 12 and FIG.
X2 = (B−Y1) × (E5 / ((E5 + E6) × 5 + (E7 + E1 + E2 + E3 + E4) × 4)).
By repeating such a series of processes until the (n-1) th day, the prediction can be corrected.

(Eighth embodiment)
The eighth embodiment of the present invention will be described.
In the present embodiment, when the target value is changed in the middle of the month, the function of reflecting the increase / decrease in the total power consumption prediction value due to the change.
This function is realized by the same processing as the total power consumption prediction correction function of the seventh embodiment.

  Here, the total power consumption prediction value for the prediction target month calculated by the power consumption prediction algorithm every month means a target value of 0% reduction target.

For example, if the target value is to reduce the total power consumption per month by 20%,
Predicted total power consumption for the prediction target day = (Predicted total power consumption for the month to which the prediction target day belongs × 0
. 8-Actual value up to the prediction target day) x (electricity ratio of the day corresponding to the prediction target day / sum of the electric energy ratio for all days from the prediction target day to the end of the month among the days belonging to the prediction target month)
It becomes.

For example, when the target value is changed to 30% reduction during the month, the result is as follows.
Predicted total power consumption for the prediction target day = (Predicted total power consumption for the month to which the prediction target day belongs × 0
. 7-Actual value up to the forecast target day) x (electricity ratio of the day corresponding to the forecast target day / sum of the power ratio for all days from the forecast target day to the end of the month belonging to the forecast target month)
Thus, when the target value is changed in the middle of the month, the increase or decrease in the predicted total power consumption due to the change can be reflected.

  Note that the processing and control described above can be performed by software processing using a CPU (Central Processing Unit) or GPU (Graphics Processing Unit), ASIC (Application Specific Integrated Circuit), or FPGA (Field Programmable Hardware). it can.

  In the above-described embodiment, the configuration and the like illustrated in the accompanying drawings are not limited to these, and can be changed as appropriate within the scope of the effects of the present invention. In addition, various modifications can be made without departing from the scope of the object of the present invention.

Each component of the present invention can be arbitrarily selected, and an invention having a selected configuration is also included in the present invention.

  In addition, a program for realizing the functions described in the present embodiment is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into a computer system and executed to execute processing of each unit. May be performed. The “computer system” here includes an OS and hardware such as peripheral devices.

  Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.

  The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory in a computer system serving as a server or a client in that case, and a program that holds a program for a certain period of time are also included. The program may be a program for realizing a part of the above-described functions, or may be a program that can realize the above-described functions in combination with a program already recorded in a computer system. At least a part of the functions may be realized by hardware such as an integrated circuit.

The present invention includes the following disclosure.
(Appendix)
(1)
Based on the total power consumption of a plurality of devices in the unit period of the previous year corresponding to the unit period to be predicted, the total power consumption prediction unit within the unit period that predicts the total power consumption of the unit period,
A day of the week coefficient calculation unit for obtaining a day of the week coefficient that is a ratio of the total power consumption of a plurality of devices on each day to the total power consumption on all days;
The day of the week coefficient calculated by the coefficient calculation unit corresponding to the day of the prediction target day by dividing the total power consumption of the unit period obtained by the total power consumption prediction unit within the unit period by the number of days included in the unit period A device control apparatus comprising: a daily total power consumption prediction unit that predicts the daily total power consumption of a plurality of devices by multiplying by.
(2)
Based on the actual power consumption values of the plurality of devices in the past T days, a total power consumption transition prediction unit for predicting the integrated power consumption transition of the plurality of devices for each time zone of the prediction target day;
The total daily power consumption calculated by the daily total power consumption transition prediction unit is matched with the total daily power consumption obtained by the total daily power consumption prediction unit. The apparatus control apparatus according to (1), further comprising an integrated power amount transition correction unit that corrects the total power consumption transition for each day.
(3)
The daily total power consumption transition prediction unit is:
An on-period determination unit that determines the on-period of the device;
The device control apparatus according to (1), further comprising: a power consumption amount prediction unit that predicts power consumption of the device by allocating on power to the on period determined by the on period determination unit.
For example, it is possible to predict power consumption with higher accuracy by applying on / off prediction for an air conditioner, eco-cute, or the like.
(4)
The apparatus is controlled or advised so that the power consumption amount becomes a preset target value based on the integrated power amount prediction value for each unit time corrected by the integrated power amount transition correction unit. The device control apparatus according to (1).
(5)
Based on the total power consumption of a plurality of devices in the unit period of the previous year corresponding to the unit period to be predicted, predicting the total power consumption in the unit period for predicting the total power consumption of the unit period,
A day of the week coefficient calculation step for obtaining a day of the week coefficient that is a ratio of the total power consumption of a plurality of devices on each day to the total power consumption on all days;
The day of the week coefficient calculated in the coefficient calculation step corresponding to the day of the prediction target day by dividing the total power consumption of the unit period determined in the unit period total power consumption prediction step by the number of days included in the unit period A device control method comprising: a daily total power consumption prediction step of predicting a daily total power consumption of a plurality of devices by multiplying.
(6)
The predicted value of the total power consumption includes the first total power consumption in the unit period immediately before the unit period of the current year, the second total power consumption in the unit period immediately before the unit period of the previous year, and the current year in the previous year. The device control apparatus according to (1), wherein a value obtained by performing an initial prediction depending on the year and month based on a ratio with the third total power consumption in the same unit period is used.
(7)
If the previous year's data exists, the actual value of the same unit period of the previous year is corrected by using the correlation between the average temperature and the total power consumption for each unit period of the previous year and the predicted temperature of the target unit period. (1) The apparatus control device according to (1), characterized in that the value is obtained by performing temperature-dependent correction for predicting the total power consumption during the prediction target unit period.

  The present invention is applicable to a device control system.

DESCRIPTION OF SYMBOLS 1 ... Device control apparatus, 11 ... Power consumption prediction part, 15 ... Equipment control part, 17 ... Control part, 17a ... Display control part, 21 ... Storage part, 21a ... Power consumption data storage part, 23 ... Display part, 25: Input unit.

Claims (5)

Based on the total power consumption of a plurality of devices in the unit period of the previous year corresponding to the unit period to be predicted, the total power consumption prediction unit within the unit period that predicts the total power consumption of the unit period,
A day of the week coefficient calculation unit for obtaining a day of the week coefficient which is a ratio of the total power consumption of a plurality of devices on each day to the total power consumption on all days;
The day of the week coefficient calculated by the coefficient calculation unit corresponding to the day of the prediction target day by dividing the total power consumption of the unit period obtained by the total power consumption prediction unit within the unit period by the number of days included in the unit period A device control apparatus comprising: a daily total power consumption prediction unit that predicts the daily total power consumption of a plurality of devices by multiplying by. A daily power consumption transition prediction unit that predicts a cumulative power consumption transition of a plurality of devices for each time zone of the prediction target day based on the actual power consumption values of the plurality of devices in the past T days;
The total daily power consumption obtained by the daily power consumption transition prediction unit is matched with the daily total power consumption obtained by the daily total power consumption prediction unit. The apparatus control apparatus according to claim 1, further comprising: an integrated power amount transition correction unit that corrects a daily power consumption amount transition. The daily power consumption transition prediction unit
An on-period determination unit that determines the on-period of the device;
The device control apparatus according to claim 1, further comprising: a power consumption amount prediction unit that predicts power consumption of the device by allocating on power to the on period determined by the on period determination unit.   The apparatus is controlled or advised so that the power consumption amount becomes a preset target value based on the integrated power amount prediction value for each unit time corrected by the integrated power amount transition correction unit. The apparatus control apparatus according to 1. Based on the total power consumption of a plurality of devices in the unit period of the previous year corresponding to the unit period to be predicted, predicting the total power consumption in the unit period for predicting the total power consumption of the unit period,
A day of the week coefficient calculation step for obtaining a day of the week coefficient that is a ratio of the total power consumption of a plurality of devices on each day to the total power consumption on all days;
The day of the week coefficient calculated in the coefficient calculation step corresponding to the day of the prediction target day by dividing the total power consumption of the unit period determined in the unit period total power consumption prediction step by the number of days included in the unit period A device control method comprising: a daily total power consumption prediction step of predicting a daily total power consumption of a plurality of devices by multiplying.

Images