JP2018157649A - Negative watt calculation method and negative watt calculation program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a negative watt calculation method and program, by which a highly reliable baseline is determined to enable effective negative watt transaction.SOLUTION: Some of demand response participants are specified as target persons, and a day in which all the target persons are targets of requested electricity saving is specified as a target day. A representative pattern of recorded power demand patterns of the target day in the target persons is specified as a target pattern. A plurality of recorded power demand patterns of the target day but not specified as targets of requested electricity saving, are specified as a reference pattern group. The difference value from the target pattern is calculated for each of the reference patterns. The reference pattern group is divided into an accepted group and an unaccepted group on the basis of the magnitude of the difference value. The maximum difference value of the accepted group is set to be less than the minimum difference value of the unaccepted group. The average pattern of the reference patterns belonging to the accepted group is specified as a baseline. The difference between the baseline and the target pattern in a target time zone for the requested electricity saving is specified as negative watt.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a calculation method and a calculation program for negative wattage, which is a reduction amount of power demand in a demand response that suppresses the peak of power demand. More specifically, the present invention relates to a method and a program for calculating negative wattage with high accuracy by appropriately determining a baseline, which is a power demand pattern when a demand response is not performed.

  Conventionally, a demand response has been proposed that suppresses the peak of power demand by giving incentives on charges and requesting power saving to some consumers. In such demand response, it is important to accurately calculate the amount of reduction in power demand due to power saving requests. The reduction in power demand is the difference in power demand between when a power saving request is made and when it is not requested. For this reason, to calculate the reduction amount, it is necessary to define a baseline that is a temporal pattern of power demand when no power saving request is made. Various methods for determining the baseline have been proposed, all of which are based on the past power consumption performance of the consumers themselves who are the target of the power saving request. The method described in Patent Document 1 is not leaked to that example.

Japanese Patent No. 595370

  However, the above-described conventional technique has a problem. The determined baseline is less reliable. Of course, in each of the conventional techniques, a technique for improving the reliability of the baseline is made (for example, [0084] to [0089] of Patent Document 1). However, the amount of data cannot be said to be abundant as long as only the customer's own past performance is used. For this reason, there was a limit in improving reliability.

  The present invention has been made to solve the above-described problems of the prior art. That is, the problem is to provide a negative wattage calculation method and a negative wattage calculation program that enable effective negative wattage transactions by determining an appropriate baseline with high reliability.

In the negative wattage calculation method according to one aspect of the present invention, a power saving request execution date and a power saving request non-execution date for each consumer belonging to a power consumer group including a demand response participant and a non-demand response participant. The power demand result pattern for each time zone of the past day including at least the power saving request implementation date is stored in the result storage unit. Then, a part of the consumers who are demand response participants are designated as the target for the calculation of negative wattage, and the day when the power saving request has been made and all of the target persons are the targets of the power saving request. A calculation target designating step for designating as a target day for negative wattage calculation, a target pattern designating step for reading out a power demand actual result pattern for the target person from the stored contents of the result storage unit, and using the representative pattern as a target pattern; A reference pattern selection step that reads a plurality of power demand performance patterns of the target day that are not the targets of the power saving request on that day from the storage contents of the result storage unit, and belongs to the reference pattern group For each reference pattern, an error value calculation step for calculating an error value with respect to the target pattern, and a reference pattern Each reference pattern belonging to the selected group is divided into an adopted group and a non-adopted group based on the calculated error value, and the maximum error value of the reference pattern belonging to the adopted group is a reference belonging to the non-adopted group. A reference pattern narrowing step that makes the pattern smaller than the minimum error value of the pattern, a baseline determination step for calculating an average pattern of each reference pattern belonging to the selected group and determining it as a baseline, and the determined baseline and target pattern And a negative watt calculation step of calculating a difference in power saving request target time zone and setting it as negative watt. Among these, at least the target pattern designation step, the reference pattern selection step, the error value calculation step, the reference pattern narrowing step, the baseline determination step, and the negative watt calculation step are performed using an arithmetic processing unit.

  The negative wattage calculation program according to another aspect of the present invention is a program for calculating the negative wattage in demand response using an arithmetic processing unit, and uses the storage contents of the above-described result storage unit. Then, the calculation target designation step, the target pattern designation step, the reference pattern selection step, the error value calculation step, the reference pattern narrowing step, the baseline determination step, The processing unit is made to perform the above-described negative wattage calculation step.

  In any of the above-described aspects, first, a target person and a target date for negative wattage calculation are specified. The target person is a part of a consumer who is a demand response participant. There may be one person. The target date is the date when the power saving request has been made in the past and all of the subjects have been subject to the power saving request. When the target person and the target date are designated, the power demand result pattern of the target day for the target person is read from the storage contents of the result storage unit, and the representative pattern is set as the target pattern. Next, a plurality of power demand performance patterns on the target day that are not the targets of the power saving request on that day are read from the storage contents of the result storage unit. These are used as a reference pattern group. Then, an error value between each reference pattern and the target pattern is calculated. Then, each reference pattern of the reference pattern group is divided into an adopted group and a non-adopted group. That is, based on the magnitude of the error value, the maximum error value of the reference pattern belonging to the adopted group is made smaller than the minimum error value of the reference pattern belonging to the non-adopted group. Then, an average pattern of each reference pattern belonging to the adopted group is calculated, and this is determined as a baseline. Then, the difference in the power saving request target time zone between the baseline and the target pattern is calculated. This is the negawatt representing all the subjects.

  In any of the above aspects, in the error value calculation step, in each reference pattern and target pattern, only the time zone excluding the power saving request target time zone, or the power saving request target time zone and a predetermined length before and after the time zone. More preferably, only the time zone excluding this time zone is the target of error value calculation. In this way, the power saving request target time zone or the power saving request target time zone and the time zones before and after the power saving request target time zone are excluded from the error value calculation target.

  According to this configuration, there is provided a negative wattage calculation method and a negative wattage calculation program that enable effective negative wattage transactions by determining an appropriate baseline with high reliability.

It is a block diagram which shows the system configuration | structure which concerns on this form. It is a graph which shows the concept of negative wattage. It is a graph which shows the example of the content of the electric power demand performance file about a demand response participant. It is a graph which shows the example of the content of the electric power demand performance file about a demand response non-participant.

  DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments embodying the present invention will be described below in detail with reference to the accompanying drawings. In this embodiment, the present invention is applied to a demand response that suppresses a peak of power demand caused by requesting power saving to some consumers in a power supply infrastructure. Furthermore, the calculation of negative wattage, which is a reduction in power demand due to a power saving request, is an application target of the present invention.

  A block diagram of a system configuration to which this embodiment is applied is shown in FIG. The system shown in FIG. 1 includes an arithmetic processing unit 1, a customer performance memory 2, and a negative watt memory 3. Thereby, the demand response in the electric power supply infrastructure 4 comprised by the electric power supply provider 5 and the consumer 6 is managed. However, in the present embodiment, description will be given with an emphasis on the negative wattage calculation part of the demand response. As shown in FIG. 2, the negative wattage is the difference obtained by subtracting the actual demand when requesting power saving by demand response from the baseline. The baseline is the power demand that is assumed when no power-saving request is made.

  In the customer performance memory 2 in FIG. 1, past power demand performance is recorded for each customer 6. Specifically, all the consumers 6 are divided into a demand response participant group (FIG. 3) and a non-demand response participant group (FIG. 4), and the past power demand record is recorded in each file. ing.

  That is, in the demand response participant's power demand record file of FIG. 3, the past power demand record pattern for each day is recorded for each ID of the consumer who is the participant. In addition, for each date recorded, a distinction is made as to whether or not that date was the date of request for power saving. However, even on the power saving request implementation date, not all demand response participants are eligible for power saving requests on that day. The power demand performance pattern subject to the power saving request is marked with a circle in the upper left in FIG. In the power demand record file of the non-participant demand response in FIG. 4, the past power demand record pattern for each day is recorded for each ID of the consumer who is a non-participant.

  The date of the recorded power demand pattern is common in FIG. 3 and FIG. The power demand performance pattern is a pattern of power demand performance for each time zone in one day. The unit of the time zone is not particularly limited, but is preferably in the range of 30 minutes to 2 hours. In this embodiment, it is assumed that it is one hour.

  Returning to FIG. 1, the arithmetic processing device 1 is provided with a target specifying function, a reference pattern selection function, an error value calculation function, a reference pattern narrowing function, a baseline determination function, and a negative wattage calculation function. These functions are of course programmed and stored in the arithmetic processing unit 1. Thereby, based on the recorded contents of the customer performance memory 2, a baseline is determined and further negative watts are calculated. The calculated negative wattage is stored in the negative wattage memory 3. The determination of the baseline and the calculation of the negative wattage are performed for each of the performance patterns stored in the demand response participant file of FIG.

Baseline determination and negative wattage calculation by the arithmetic processing unit 1 are performed in the following procedure.
1. Designation of target power demand performance pattern ↓
2. Selection of reference pattern group that is the basis of the baseline ↓
3. Calculation of error value for each reference pattern ↓
4). Narrow down the reference pattern group ↓
5. Baseline decision based on narrowed-down reference pattern group ↓
6). Calculation of negative wattage based on the determined baseline

  First, “1. Specification of target power demand record” will be described. This designation is to designate the target person and the target date for the negative wattage calculation. Naturally, the target person is a part of the consumer listed in the file of the demand response participant shown in FIG. There may be one person or a plurality. The target date is the power saving request implementation date (the day of “with DR” in FIG. 3), which is one day on which all designated target persons are targeted for the power saving request. In FIG. 3, it can be determined whether or not each designated target person is the target of a power saving request on that day by whether or not a circle is added to the upper left of each pattern. Although not shown in FIG. 3, the time zone for which the power saving request is made is also recorded in the participant file.

  If the target person and the target date are determined, the power demand pattern to be calculated for the negative wattage is determined. Therefore, the pattern is read from the customer performance memory 2. The average and other representative patterns are “target patterns”. When there is one target person, the actual power demand pattern is directly the “target pattern”. The designation of the target person and the target date here may be performed manually by the operator, or may be automatically specified by the arithmetic processing device 1. In general, it is desirable that all the power demand performance patterns in FIG. 3 that are circled are sequentially set as target patterns. That is, it is desirable to determine the baseline and calculate the negative wattage for all target patterns. Of course, if the baseline has already been determined and the negative wattage has already been calculated, there is no need to set it as the target pattern again.

  Next, “2. Selection of a reference pattern group as a base of a baseline” will be described. In this step, a power demand record pattern that is a basis for determining a baseline for the “target pattern” specified above is selected. A power demand performance pattern related to the same date as the target pattern is selected, and it is not selected as a power saving request target. The reason why the date with the same date as the target pattern is selected is that the day attribute (day of the week, season, presence / absence / type of singular event) is the same as the date of the target pattern as the basis of the baseline. It is. The reason why the power saving request is not selected is that the meaning of the baseline is the actual power demand pattern assumed when the power saving request is not made.

  For this reason, in this step, a plurality of power demand performance patterns for the corresponding day are selected from the file of non-demand response participants in FIG. All the corresponding power demand performance patterns may be selected, or may be selected after being narrowed down to some extent. Even when narrowing down the selection, it is desirable that the number of power demand pattern to be selected is about 10,000. This is to eliminate the influence of individual customer's personality as much as possible. The selected power demand record pattern is read out and becomes a “reference pattern”.

Random sampling may be used for narrowing down. This is because, in a separate refinement step, which will be described later, it is narrowed down to consumer patterns having attributes (family form, residence form, etc.) close to the target person. However, it is not prohibited to perform the narrowing down based on the customer's attributes at this step. Of course, for this purpose, it is necessary to add some data for classifying consumers by their attributes in FIGS. Further, the reference pattern may be selected not only from the non-participant file of FIG. 4 but also from the participant file of FIG. However, the patterns selected from FIG. 3 are limited to those of the corresponding day and not marked with the above-described circle.

  As a result, a “target pattern” and a “reference pattern” group corresponding to the “target pattern” are arranged. Next, “3. Calculation of error value for each reference pattern” will be described. In this step, a rating is given to how similar each selected reference pattern is to the target pattern.

  Therefore, each of the target pattern and the reference pattern is represented by a vector. That is, if the target pattern is X and the reference pattern is Yi (i is a number for individually identifying a plurality of reference patterns), X and Yi are expressed by the following equations. Here, “j” is a time zone number when the daily power demand record pattern is divided into time zones. The component of each vector is a power demand value in each time zone.

  Then, the error value Ei between X and Yi can be calculated by the following equation. “Σ” in this equation means the sum of “j”. “M” is not particularly limited as long as it is a positive number of 1 or more, but “2” is usually sufficient. That is, the error value Ei is the absolute value of the difference between the power demand values for each time zone between the target pattern and the reference pattern or the sum of the power values thereof.

  The error value Ei calculated in this manner indicates that the smaller the value is, the more similar the reference pattern Yi is to the target pattern X. Conversely, when the error value Ei is large, it indicates that the reference pattern Yi is not similar to the target pattern X. That is, it can be said that the similarity between the reference pattern Yi and the target pattern X is rated according to the magnitude of the error value Ei. Further, the fact that the reference pattern Yi is similar to the target pattern X is considered that the consumer attribute of the reference pattern Yi is close to the consumer attribute of the target pattern X. Of course, the calculation of the error value Ei is performed for all “i”.

  Here, in calculating the above error value Ei, it is better to exclude “j” corresponding to the time zone for which a power saving request is made by demand response from the sum of “Σ”. In the time period, the target pattern X is affected by the power saving request, but the reference pattern Yi is not received. Therefore, even if the consumer attribute of the reference pattern Yi is close to the consumer attribute of the target pattern X in this time zone, the power demand value is considered to be greatly different between the two. For this reason, the power demand value in this time zone is not very suitable for evaluating the similarity between the reference pattern Yi and the target pattern X. Therefore, the similarity can be more appropriately evaluated by calculating the error value Ei outside this time zone.

  More specifically, not only the target time zone for requesting power saving, but also a certain time zone before and after that (for example, about 1 hour may be determined in advance) may be further excluded from the calculation of the error value Ei. This is because, in the time zone before and after the target time zone for requesting power saving, there is a possibility that the consumer who was the target of the power saving request was performing different power consumption behavior.

  Next, “4. Refinement of reference pattern group” will be described. In this step, the “reference pattern” group for which the error value Ei is calculated is narrowed down to reduce the number. That is, the reference pattern group is divided into an adopted group that is actually used for determining the baseline and a non-adopted group that is not actually used. Of course, the reference pattern put into the adopted group has a small error value Ei, and the reference pattern put into the non-adopted group has a large error value Ei. That is, the maximum value of the error value Ei in the adopted group is made smaller than the minimum value of the error value Ei in the non-adopted group. As a result, the reference pattern group is narrowed down to only those of customers who are attributed to those of the target pattern. By performing this narrowing down, the accuracy of the baseline determined in the next step is improved.

  From the above, when the entire reference pattern group is arranged in ascending order according to the magnitude of the error value Ei, up to a certain number of reference patterns from the smallest error value Ei belong to the selected group. Here, there are various determination methods for how many reference patterns are included in the adopted group, and there is no particular limitation. However, preferably, the number of reference patterns belonging to the adopted group is preferably about 0.1 to 0.3 times as a ratio to the total number of reference pattern groups. The number itself is desirably about 1000 or more. If the number is too small, there is a possibility that the influence of the individuality of each consumer cannot be erased. On the other hand, if the number is too large, a reference pattern by a consumer having an attribute that is not similar to the consumer attribute of the target pattern is mixed. Both are factors that lower the accuracy of the baseline determined in the next step.

  For example, the number of reference patterns belonging to the selected group can be determined in advance. Alternatively, the number of reference patterns belonging to the adopted group can be defined as a ratio to the total number of reference pattern groups. It is also possible to set a threshold value for the error value Ei and put a reference pattern having an error value Ei equal to or less than the threshold value into the selection group. The threshold value may be determined in advance or may be an average value or other representative value for the entire error value Ei actually calculated. Furthermore, a method is also possible in which the entire reference pattern group is arranged in ascending order according to the magnitude of the error value Ei, and the boundary between the adjacent error values Ei is the largest. Some of these techniques can also be used in combination.

  Next, “5. Determination of baseline based on narrowed-down reference pattern group” will be described. In this step, an average of the reference patterns Yi belonging to the selected group is taken and used as a baseline. That is, the power demand value for each time zone of each reference pattern Yi is averaged for each time zone. A power demand pattern constituted by the average power demand value for each time period obtained in this way is determined as a baseline. That is, the baseline vector display B and its components are expressed as follows: Here, “Σ” in the equation means the sum of reference patterns belonging to the adopted group among “i”. “M” is the number of reference patterns belonging to the adopted group.

  It should be noted that in the error value calculation step described above, regarding the calculation of the error value Ei, it is possible to exclude the target time zone for requesting power saving, or the target time zone and a predetermined time zone before and after the target time zone. It was. Even when the error value Ei is calculated as described above, in this step, the baseline B is determined for all time zones including the target time zone for requesting power saving. This is because the baseline B is required only in the target time zone of the power saving request in order to calculate the negative power in the next step.

  Next, “6. Calculation of negative wattage based on the determined baseline” will be described. In this step, the difference between the target pattern X and the baseline B is obtained, and the result is defined as negawatt W. Negawatt W is given by: Here, “Σ” in the equation means the sum of “j” corresponding to the time zone for which the power saving request is made.

  The negative wattage W calculated in this way is a negative value, and indicates the wattage in the target pattern X that can be evaluated as having been able to save power due to a request for power saving. The calculated negative watt W is stored in the negative watt memory 3 together with information for identifying the corresponding target date and target customer. The negative wattage W calculated as described above in this embodiment is extremely accurate. The reason why the accuracy is high is that the accuracy of the baseline B is high. The reason why the accuracy of the baseline B is high is that it is determined based on a reference pattern group appropriately narrowed down based on the error value Ei. That is, the baseline B is determined based only on the reference pattern by the consumer whose attributes are similar to those of the consumer of the target pattern X while eliminating the influence of the individuality of each consumer. When there are a plurality of subjects, a negative watt representing the entire subject is calculated.

  As described above in detail, according to the present embodiment, the determination of the baseline for the calculation of negative wattage is a power demand performance pattern on the calculation target date, and is not subject to the power saving request on the target date. It is based on multiple things. In other words, although the power demand pattern is different from the calculation target consumer, the target date is the same and the number of cases is sufficiently large. Furthermore, by narrowing down by error values, the power demand performance pattern by consumers whose attributes are different from those of the consumers to be calculated is excluded from the determination of the baseline. The highly reliable baseline determined in this way is used for the calculation of negative wattage.

  For this reason, by using the negative wattage calculated according to the present embodiment, it is possible to smoothly carry out the negative wattage transaction for demand response. Also, it is desirable to calculate the negative wattage according to this embodiment for each of the many power demand performance patterns that are the targets of the power saving request. By doing so, the prediction accuracy of the total amount of negative watts that can be expected in the future when making a power saving request is improved, and the target customers of the power saving request can be selected more appropriately. In this way, a negative wattage calculation method and a negative wattage calculation program that enable effective negative wattage trading by determining an appropriate baseline with high reliability are realized.

  Note that this embodiment is merely an example, and does not limit the present invention. Therefore, the present invention can naturally be improved and modified in various ways without departing from the gist thereof.

  For example, in the present embodiment, in the description of FIG. 3, even on the power saving request implementation date, all demand response participants are not necessarily targeted for power saving requests on that day. However, the present invention is not limited to such a demand response system, and the present invention can be applied to a demand response system in which all demand response participants are targeted for a power saving request on the power saving request implementation date. In that case, “selection of reference pattern group” is performed only from the non-participant file of FIG.

Further, in the present embodiment, the calculation of the negative wattage W after the baseline determination is performed only for the time zone for which the power saving request is made. However, the present invention is not limited to this, and it may be performed for a target time zone for requesting power saving and a predetermined time zone before and after that. As described above, power consumption behavior may have been different from the normal time period before and after the target time period for requesting power saving, and this includes the amount of true power saving resulting from the power saving request. Because it can also be considered. 3 and 4, the present invention can be implemented even if only the power saving request execution date is recorded.

[Preliminary claim]
A negative wattage calculation device for calculating negative wattage in demand response,
For individual consumers belonging to the group of power consumers including demand response participants and non-demand response participants, a past day including at least the power saving request execution date among the power saving request execution date and the power saving request non-execution date A performance storage unit for storing power demand performance patterns for each time zone;
A part of consumers who are demand response participants are designated as the target of the calculation of negative power, and one day of the previous power saving request implementation date when all of the target persons are the target of the power saving request. A calculation target designating part that is designated as a calculation target date;
A target pattern designating unit that reads out the power demand result pattern of the target date for the target person from the stored contents of the result storage unit and uses the representative pattern as a target pattern;
A reference pattern selection unit that reads out a plurality of power demand performance patterns of the target day that are not the target of the power saving request on the day from the stored contents of the result storage unit,
An error value calculation unit for calculating an error value between each of the reference patterns belonging to the reference pattern group and the target pattern;
Each reference pattern belonging to the reference pattern group is divided into an adopted group and a non-adopted group based on the magnitude of the calculated error value, and the maximum error value of the reference pattern belonging to the adopted group is the non-adopted group. A reference pattern narrowing section that makes the reference pattern smaller than the minimum error value of the reference pattern belonging to
A baseline determination unit that calculates an average pattern of each reference pattern belonging to the adopted group and determines a baseline;
A negative wattage calculation device, comprising: a negative wattage calculation unit that calculates a difference in a power saving request target time zone between the determined baseline and the target pattern and sets the negative wattage.

DESCRIPTION OF SYMBOLS 1 Arithmetic processor 2 Customer results memory 3 Negawatt memory

Claims (4)

For individual consumers belonging to the group of power consumers including demand response participants and non-demand response participants, a past day including at least the power saving request execution date among the power saving request execution date and the power saving request non-execution date Store the power demand performance pattern for each time zone in the performance storage section,
A part of consumers who are demand response participants are designated as the target of the calculation of negative power, and one day of the previous power saving request implementation date when all of the target persons are the target of the power saving request. A calculation target designating step that is designated as a calculation target date;
A target pattern designating step of reading out the power demand result pattern of the target date of the target person from the stored content of the result storage unit and using the representative pattern as a target pattern;
A reference pattern selection step of reading out a plurality of power demand performance patterns of the target day that are not the targets of the power saving request on the day from the stored contents of the result storage unit, and making a reference pattern group;
An error value calculating step for calculating an error value between each of the reference patterns belonging to the reference pattern group and the target pattern;
Each reference pattern belonging to the reference pattern group is divided into an adopted group and a non-adopted group based on the magnitude of the calculated error value, and the maximum error value of the reference pattern belonging to the adopted group is the non-adopted group. A reference pattern narrowing step that makes the reference pattern smaller than the minimum error value of the reference pattern belonging to
A baseline determination step of calculating an average pattern of each reference pattern belonging to the adopted group and determining a baseline;
A negative wattage calculation method comprising: calculating a difference in a power saving request target time zone between the determined baseline and the target pattern to obtain a negative wattage. The negative wattage calculation method according to claim 1, wherein in the error value calculation step,
In each reference pattern and the target pattern,
Only the time zone excluding the time zone for requesting power saving, or
Only the time zone excluding the time zone for requesting power saving and the time zone of a predetermined length before and after it,
Is a method for calculating a negative wattage, which is an error value calculation target. A negative wattage calculation program for calculating negative wattage in demand response using an arithmetic processing unit,
For individual consumers belonging to the group of power consumers including demand response participants and non-demand response participants, a past day including at least the power saving request execution date among the power saving request execution date and the power saving request non-execution date Using the memory contents of the record storage unit that stores the power demand pattern for each time period,
A part of consumers who are demand response participants are designated as the target of the calculation of negative power, and one day of the previous power saving request implementation date when all of the target persons are the target of the power saving request. A calculation target designating step that is designated as a calculation target date;
A target pattern designating step of reading out the power demand result pattern of the target date of the target person from the stored contents of the result storage unit and using the representative pattern as a target pattern;
A reference pattern selection step of reading out a plurality of power demand performance patterns of the target day that are not the targets of the power saving request on the day from the stored contents of the result storage unit, and making a reference pattern group;
An error value calculating step for calculating an error value between each of the reference patterns belonging to the reference pattern group and the target pattern;
Each reference pattern belonging to the reference pattern group is divided into an adopted group and a non-adopted group based on the magnitude of the calculated error value, and the maximum error value of the reference pattern belonging to the adopted group is the non-adopted group. A reference pattern narrowing step that makes the reference pattern smaller than the minimum error value of the reference pattern belonging to
A baseline determination step of calculating an average pattern of each reference pattern belonging to the adopted group and determining a baseline;
A negative wattage calculation program for causing an arithmetic processing unit to perform a negative watt calculation step of calculating a difference in a power saving request target time zone between a determined baseline and the target pattern to make a negative wattage. The negative wattage calculation program according to claim 3, wherein in the error value calculation step,
In each reference pattern and the target pattern,
Only the time zone excluding the time zone for requesting power saving, or
Only the time zone excluding the time zone for requesting power saving and the time zone of a predetermined length before and after it,
Is a negative wattage calculation program characterized in that the error value is calculated.

Images