JP5287478B2 - Power energy monitoring system - Google Patents

Description

  The present invention relates to a power energy monitoring system that monitors the amount of power energy used by consumers, and particularly relates to the calculation of a target value and a predicted value of the amount of power energy used by each customer.

  There is an increasing need for functions that support energy conservation activities of power consumers. In conducting energy saving activities, it is first decided to grasp the actual state of energy use. By this grasping, it is possible to identify energy saving improvement points and energy saving activities.

  The monitoring system for the amount of electric energy used is measured by the power company on the supply side and the PPS (specific electric utility) etc., and the energy consumption exceeds the preset value. When trying to do so, messages and information that encourage consumers to save energy are sent.

In addition, as an energy-saving system, customers themselves monitor the current value as a real-time trend for the purpose of reducing power consumption and CO ₂ emissions. From this monitoring, the daily and monthly power consumption is displayed as a bar graph for each fiscal year. Some display (for example, refer to Patent Document 1).

  As another energy-saving system, a method of contracting an electricity charge between a large power consumer and an electric power company up to a maximum value (maximum demand) for an average power consumption for a predetermined time period, for example, every 30 minutes or 60 minutes There is. That is, if the demand value increases, the contract power increases, and even if the power consumption is the same, the electricity bill increases as the contract power increases. For this reason, it is effective to keep contract power low to reduce energy costs, and it is necessary to reduce power demand to reduce the amount of power used. Yes.

  Power demand monitoring calculates the power demand from the power amount pulse data at a fixed period, and when the predicted demand obtained as a result of the calculation is expected to exceed the target demand, or an alarm setting in which the current demand is preset An alarm is displayed when the value deviates. In addition, power demand data, prediction results, and target power are displayed in a trend graph format. Furthermore, there is also a control that performs control by turning on / off the load so that the predicted value of the demand value does not exceed the contract power (see, for example, Patent Document 2).

JP 2006-309325 A Japanese Patent Laid-Open No. 08-0663132

  The conventional power demand monitoring system measures the amount of current power energy used, predicts the amount of power energy used 30 or 60 minutes after the current point in time, and calculates the current value, predicted value, and contract power. It is displayed in a trend graph format, and an alarm is displayed when the predicted value deviates from preset contract power or energy saving target power.

  In this monitoring system, the prediction of the amount of power energy used is limited to a short time such as 30 minutes or 60 minutes later, and is not predicted or monitored over a long period of time such as a day or month. Not. Moreover, since it is difficult to predict over a long period of time, it is difficult to accurately set the contract power and the energy saving target power, resulting in high contract power or a reduced energy saving effect.

  Moreover, since the customer side cannot grasp the actual use of energy after a long period of time, it cannot make an operation plan for a long period of time. On the other hand, the power company predicts short-term and long-term power demand for the entire region, but it cannot predict long-term power demand for each customer, which may hinder power supply and demand and distribution plans. is there.

  An object of the present invention is to provide a power energy monitoring system that can accurately and accurately monitor power energy consumption by obtaining predicted and target values of power energy usage over a long period of time such as daily or monthly. There is to do.

  In order to solve the above-mentioned problem, the present invention obtains a target value (target pattern) of the daily and monthly power energy usage from the actual value as the past power energy usage, and the current power energy usage and From the present value and the past actual value, the predicted value (prediction pattern) of the daily and monthly power energy consumption is obtained, and the day type (weekday / holiday / holiday) is used to calculate the daily value of these target pattern and prediction pattern. Use the actual values that are the same for the weekly composition for the monthly calculation, and further narrow down and average these actual values, and target values, predicted values, and current The power energy is monitored from the value and the actual value, and it features the following monitoring system.

(1) A power energy monitoring system that supports consumer energy conservation activities by monitoring consumer energy consumption.
A database that sequentially stores the current value and past actual value of customer's power energy usage collected at regular intervals;
A target value calculation unit that obtains a target value (target pattern) of power energy usage in units of days and months based on actual values in a sample period with power energy usage stored in the database;
A predicted value calculation unit for obtaining a predicted value (prediction pattern) of power consumption in daily and monthly units from a current value and a past actual value of power energy usage stored in the database;
From the current value and the past actual value of power energy usage stored in the database, a yearly predicted value calculation unit that obtains a predicted value of yearly power energy usage as a monthly prediction pattern;
A trend graph display unit for displaying the current value and actual value of power energy consumption stored in the database, the target value obtained by the target value calculation unit and the predicted value obtained by the predicted value calculation unit in a trend graph format; ,
An alarm reporting unit that issues an alarm when the predicted value calculated by the predicted value calculating unit deviates from the target value calculated by the target value calculating unit ;
The yearly predicted value calculation unit obtains the past actual value of the power energy consumption stored in the database every month, calculates the average value of the monthly actual value for each year, and The average slope of each month is calculated, and it is determined whether the monthly average slope is increasing or decreasing. If the monthly average slope is increasing, the monthly average slope is added to the same month value of the previous year. If the average slope for each month is decreasing, subtract the average slope for each month from the previous year's same-month value, and if there are both increasing and decreasing trends in the monthly average slope, The average value for each and the average slope for each month are added or subtracted to obtain a predicted value of annual energy consumption,
The trend graph display unit is a graph format in which the past actual pattern in units of months, the prediction pattern in units of months, and the target pattern in units of months acquired automatically or manually are combined in units of years. It is characterized by being displayed .

  (2) The daily target pattern calculation by the target value calculation unit and the calculation of the daily prediction pattern by the prediction value calculation unit comprise means for using the actual value combined with the day type. To do.

  (3) The monthly target pattern calculation by the target value calculation unit and the monthly prediction pattern calculation by the prediction value calculation unit comprise means for using the actual values having the same weekly composition. To do.

  (4) The target value calculation unit and the predicted value calculation unit include means for performing calculation by narrowing down and averaging the actual values.

  (5) The target value calculation unit includes means for calculating a ratio between the value of the final time of the calculated target pattern and the target value and correcting the deviation from the target value by multiplying the calculated target pattern by the ratio. It is characterized by that.

  (6) The predicted value calculation unit includes means for obtaining a difference between the calculated prediction pattern and the current value, and correcting a deviation from the current value by adding or subtracting the difference to the calculated prediction pattern. It is characterized by.

  As described above, according to the present invention, the target value (target pattern) of the daily and monthly power energy usage is obtained from the actual value that is the past power energy usage, and the current power energy usage is obtained. The estimated value (prediction pattern) of daily and monthly power energy usage is calculated from the actual value and the past actual value, and the day type (weekday / holiday / holiday / holiday) is used to calculate the target pattern and forecast pattern on a daily basis. For the monthly calculation, use the actual values with the same weekly composition, further narrow down and average these actual values, and use these target values, predicted values, and current values. Since power energy is monitored, the predicted and target values of power energy usage over a long period of time, such as daily or monthly, are obtained with high accuracy, and the power energy is accurately determined. It can be seen.

  In addition, the target pattern and the predicted pattern can be corrected to a high-accuracy pattern by correcting the target pattern and the predicted pattern to the reference target value and the current value, and reliable monitoring can be performed.

  In addition, as a grasp of the actual state of energy use, long-term (yearly) target management can be performed with respect to the amount of energy used (electric power), and awareness of energy-saving activities can be improved.

The principal part block diagram of the electric power energy monitoring system which shows embodiment of this invention. Calculation example of forecast values from actual values in daily and monthly units. The relationship diagram of a target pattern and a prediction pattern. Processing flow of daily target value or forecast value. Monthly target value or forecast value processing flow. Calculation example of daily target pattern. Calculation example of daily forecast pattern. Example of daily trend graph display. Monthly trend graph display example. An example of daily target pattern calculation deviation. An example of daily forecast pattern calculation deviation. An example of correcting a target pattern by a ratio. The example of correction | amendment of the prediction pattern by a difference. The principal part block diagram of the electric power energy monitoring system which shows embodiment of this invention. An example of past actual values of electric energy consumption. Calculation example of annual target pattern and forecast pattern. Calculation process flow of annual forecast pattern. Example of predictive pattern calculation by year when actual data is decreasing. Example of predictive pattern calculation for each year when actual data increases. Example of predictive pattern calculation by year when actual data increases / decreases. Graph format display example of actual value, predicted value, and target value of power energy usage.

(Embodiment 1)
(1) System Overview FIG. 1 is a main part configuration diagram of a power energy monitoring system showing an embodiment of the present invention, and each part constructs a system by computer resources and software using the same.

  The power energy usage collection unit 10 uses the monitoring information collection function of the existing remote monitoring and control system (TC), and the power received by each business office from a power company, etc., is constant as the power energy usage (digital amount). Collect at periodic intervals.

  The database (DB) 20 sequentially stores the power energy usage collected by the power energy usage collecting unit 10.

  The target value calculation unit 30 sets the target value (target power) of the daily and monthly power energy usage based on the actual value in the sample period with the power energy usage stored in the database 20 in the factory, the management center, etc. It is calculated for each office and set as the target value for each office.

  The predicted value calculation unit 40 obtains a predicted value of the daily and monthly power energy usage for each business establishment from the current value and past actual value of the power energy usage stored in the database 20.

  The trend graph display unit 50 reads the electric energy consumption amount stored in the database 20 as a past actual value, the actual value, the target value obtained by the target value calculating unit 30, and the predicted value obtained by the predicted value calculating unit 40. Is displayed in the trend graph format.

  The alarm notification unit 60 issues an alarm when the predicted value obtained by the predicted value calculation unit 40 deviates from the target value obtained by the target value calculation unit 30.

(2) Determination of target pattern and prediction pattern (a) Outline of processing Sample data is collected under the following conditions based on past actual values (daily / monthly report data) stored in the database 20, and further narrowed down by narrowing conditions After performing the above, an average value is calculated as a predicted value.

  The sample data is collected from the past actual values of the power energy usage, which is a monitoring item, by extracting the actual values for the day or month that match the following conditions and using them as sample data.

  ・ As shown in the example in Fig. 2 (a), the daily performance is the daily performance of the same type (weekdays / holidays / holidays / holidays, etc.) as the current day according to the calendar setting, from the day before the forecast date. All data. However, up to the upper limit of the number of data (system definition).

  -As shown in the example of FIG. 2 (b), the monthly performance values are all the monthly performance data that is the same as the weekly composition of the current month in all monthly reports from the month preceding the forecast month. However, up to the upper limit of the number of data (system definition).

  In calculating the target pattern and the predicted pattern, the extracted sample data is narrowed down by the target power amount ± m%, and then the average value is calculated to obtain the target pattern. Further, after the extracted sample data is narrowed down by the current cumulative value ± m%, an average value is calculated and set as a prediction pattern for the corresponding day or month.

  FIG. 3 shows the relationship between the target pattern and the predicted pattern based on daily and monthly narrowing.

  Note that ± m% of the narrowing-down conditions can be set for each monitoring item as the target power amount prediction range on the setting screen.

  In addition, at the time of the daily change, the target electric energy of the previous day / previous month is taken over as the target electric energy of the current day / current month at the time of monthly change. However, in the daily target power amount, the target power amount of the same type (weekday / holiday / singular day 1 / singular day 2) as that day is assumed.

  In calculating the average value, in order to thin out the special data, the value of the pattern minimum value + n% and the value of the maximum value−n% are excluded (± n% is defined as a system definition).

(B) Processing Flow FIG. 4 shows the processing flow of the daily target value or predicted value. First, the day type (weekday / holiday / holiday / holiday, etc.) of the data read from the database is acquired (S1). If this day type is the same day as the current day (S2), the past data of the same day ( 24 hours) is acquired (S3), and when the target data is within a range of ± m% for each acquired data (S4), data acquisition is repeated until the acquired data reaches a predetermined value (S5).

  When a predetermined number of past data is acquired by the above processes (S1) to (S5), the maximum value and the minimum value are acquired from the acquisition data for each time (S6), and the minimum value + n% of the acquisition data The data and the data of the maximum value −n% are excluded (S7), an average value is calculated from the remaining data (S8), and these processes are performed for 24 hours (S9). Through these processes (S6) to (S9), a daily target value pattern or predicted value pattern is obtained.

  Note that the determination of the target data in the process (S4) determines whether the 24-hour data is the target value ± m% at the time of the daily target pattern calculation. Further, when calculating the daily forecast pattern, it is determined whether or not past data at the same time as the current time is current time data ± m%.

  FIG. 5 shows a processing flow of monthly target values or predicted values. First, the day of the week before the current day is acquired (S11), and it is determined whether this is the same day of the week (S12). If the day is different, the day of the week before the acquisition date is acquired (S13). By repeating these processes (S12) and (S13), the first day of the same week structure near one month ago is searched. The past data for one month (the number of days of the current month) is acquired from the searched same day of the week (S14), and when the acquired target data is within ± m% range (S15), the data is acquired until the acquired data reaches a predetermined value. The acquisition is repeated (S16).

  When a predetermined number of past data is acquired by the above processes (S11) to (S16), the maximum value and the minimum value are acquired from the acquired data for each day (S17), and the minimum value + n% of the acquired data Then, the data of the maximum value -n% is excluded (S18), the average value is calculated from the remaining data (S19), and these processes are performed for the number of days in the current month (S20). Through these processes (S17) to (S20), a monthly target value pattern or predicted value pattern is obtained.

  The first day search of the same week structure in the process (S13) is performed, for example, when the calculated first day (1st) of the current month (for 31 days) is Tuesday and the first day of the previous month (1st) is Wednesday. The date of Wednesday of the previous week is acquired from that day as the starting point. Then, the data for the number of the current month from the acquired date is used as the previous month data.

  Further, the determination of the target data in the process (S15) is to determine whether or not the last date data of the month is the target value ± m% at the time of calculating the monthly target pattern. Further, at the time of calculating the monthly prediction pattern, it is determined whether or not past data that is the same as the current date is current date data ± m%.

(C) Specific pattern calculation example As shown in FIG. 6, the calculation of the daily target pattern includes the past performance data patterns P1 to P5 and P2 ′, P4 ′, P3 ′, and P3 ″. In this case, assuming that the target power amount is 100 kWh, the actual data entering the narrowing-down condition (100 kWh ± m%) at 24:00 is P2 to P4, P2 ′, P4 ′, P3 ′, and P3 ″. Based on the narrowed down result data, each of 1 o'clock to 24 o'clock is averaged.

  When averaging is performed, the minimum value + n% and the maximum value -n% are further excluded in order to thin out the special data. As a result, P2, P2 ', P4, and P4' are excluded.

  Accordingly, when a daily target pattern with a target power amount of 100 kWh is calculated from the actual pattern, the patterns P3, P3 ', and P3 "of the actual data are averaged.

  The calculation of the daily forecast pattern is based on the amount of power at 24 o'clock in the above target pattern, whereas the forecast pattern for the day is narrowed down based on the amount of power at the current time. The future pattern of the pattern is used as the predicted pattern.

  For example, as shown in FIG. 7, if the current time is 18:00 and the actual data is 80 kWh, the actual data that falls within the narrowing-down condition (80 kWh ± m%) is P1 to P3, P2 ′, P3 ′, and P3 ″.

  Based on this narrowed-down performance data, each of 18:00 to 24:00 is averaged. When averaging is performed, the minimum value + n% and the maximum value -n% are further excluded to thin out the special data, and the results P1 and P3 'are excluded.

  Accordingly, when a predicted pattern with a predicted power consumption of 80 kWh is calculated from the above-mentioned actual pattern, the actual data P2, P2 ', P3, and P3' 'are averaged patterns.

  In addition, although the above shows the case of the daily target pattern and the prediction pattern, the monthly pattern calculation is the same.

(3) Trend graph display of current value, prediction result, and target value FIG. 8 shows an example of a daily trend graph display, and FIG. 9 shows an example of a monthly trend graph display.

  The daily display shows the graph and real number display of the previous day monitoring and the same day monitoring, the previous day monitoring displays the target power amount A of the previous day and the current power amount B of the previous day, and the current day monitoring displays the target power amount C of the current day. And the predicted power amount D of the day are displayed.

  Similarly, for the monthly display, the previous month monitoring and current month monitoring graphs and real numbers are displayed, the previous month monitoring displays the previous month target power E and the previous month current power F, and the current month monitoring displays the current month target. The power amount G and the predicted power amount H for the current month are displayed.

  The prediction width (0% to 100%) during the real number display of the target power amount designates the target power amount used in the sample data narrowing condition and the width of the current power amount. For example, if 10% is designated, data that coincides with the range of the target power amount ± 10% and the current power amount ± 10% is set as sample data. The prediction mask period is a period in which period prediction is not performed, and is used when the actual data is not complete and correct prediction cannot be performed.

  The remaining power amount displays the difference between the current power amount and the predicted power amount. In the remaining period, the remaining time or the number of days until the monitoring end of the current day or the current month is displayed. The previous day (previous month) electric energy indicates the electric energy of the previous day or the previous month.

(4) Alarm issuance Alarm is issued when the predicted value deviates from a preset target value.

  Alerts when the monthly cumulative total (the cumulative total of the target month) exceeds each target value due to the calculated forecast pattern at the time of day for daily monitoring and at the time of daily monitoring for monthly monitoring. Print a message with output.

  Note that even if the predicted pattern for the halfway period (from 0 o'clock to 23:00) in daily units exceeds the target pattern, no alarm is output unless the final cumulative value exceeds the target value. The same applies to the intermediate period in the monthly unit (1 day to the last day of the month). In addition, at the next forecast timing or target pattern update by changing the setting value, the daily final cumulative value (24 hour cumulative value) or monthly final cumulative value (corresponding month final day cumulative value) of the prediction pattern It returns when it falls below.

  When the alarm is issued, the alarm status is also displayed on the monitoring screen of FIG. 8 or FIG. Of these, when an alarm is occurring, it is displayed while the alarm is occurring. During initialization, it is displayed when the system is stopped and restarted during monitoring, and is cleared when one monitoring cycle elapses. Since monitoring starts from the middle of the monitoring cycle, it indicates that the data is not accurate.

  The target power amount missing measurement is displayed when there is a period that cannot be calculated due to the past actual data missing measurement or the like when the target pattern is calculated. The predicted power amount missing measurement is displayed when there is a period that cannot be calculated due to the past actual data missing measurement or the like when the predicted pattern is calculated.

  Therefore, according to the present embodiment, it is possible to accurately predict power energy usage over a long period of time, such as a daily unit or a monthly unit, and to obtain a target value with high accuracy, thereby accurately monitoring the electric energy.

  Furthermore, as one of the grasps of the actual situation of energy use, target management for energy use (power consumption) can be performed, and awareness of energy-saving activities can be improved.

  Further, by issuing an alarm when the predicted value deviates from the target value, it is possible to take measures before the energy usage exceeds the target value.

(Embodiment 2)
In the first embodiment, the predicted value calculation method collects an approximate pattern of a target value serving as a reference or a current value from past actual values and averages them as a target pattern and a predicted pattern. Deviation from the target value or current value (decrease in accuracy) occurs.

  For example, in the daily target pattern calculation shown in FIG. 10, when the target pattern of the target power amount 100 kWh is calculated by narrowing down from the performance patterns P1 to P5 and P2 ′, P4 ′, P3 ′, P3 ″, The result is a dotted line pattern Pt obtained by averaging the patterns P3, P3 ′, and P3 ″ of the result data, and the target power amount 100 kWh and the value of the target pattern at 24 o'clock do not match.

  Further, in the calculation of the daily prediction pattern shown in FIG. 11, when the prediction pattern of the current power amount 80 kWh is calculated from the actual pattern, the dotted line pattern obtained by averaging the actual data P2, P2 ′, P3, and P3 ″ Pf, and the current power amount of 80 kWh does not match the predicted pattern value at 18:00.

  In the present embodiment, the following correction processing function is provided to match the target value or the current value based on the target pattern and the prediction pattern.

(Correction process 1)
The ratio between the calculated target pattern and the target value is obtained, and the deviation from the target value is corrected by multiplying the target pattern.

  For example, as shown in FIG. 12, when a target pattern with a target power amount of 100 kWh is calculated from a past performance pattern and a dotted line pattern P10 is obtained, the following correction processing is performed.

  The ratio α between the target power amount and the target pattern P10 at 24:00 (final time) is calculated.

  The data 1 o'clock to 24:00 of the target pattern P10 is multiplied by the ratio α to obtain a corrected target pattern P10 '.

  Specifically, when the 24-hour value of the target pattern calculated for the target power amount of 100 kWh is 125 kWh, the ratio α is 100 ÷ 125 = 0.8, and the target pattern (1 By multiplying by ˜23: 00, the pattern is corrected within the range of 0:00 (0 kWh) to 24:00 (100 kWh).

(Correction process 2)
A difference between the calculated prediction pattern and the current value is obtained and added to (or subtracted from) the prediction pattern to correct a deviation from the current value.

  For example, as shown in FIG. 13, when a predicted pattern of the current power amount of 80 kWh is calculated from the past performance pattern P10 and a dotted line pattern P20 is obtained, the following correction processing is performed.

  The difference β between the 18:00 data of the performance pattern P10 and the 18:00 data of the predicted pattern P20 is obtained.

  The difference β is added (or subtracted) to the 18:00 to 24:00 data of the predicted pattern P20 to obtain a corrected predicted pattern P20 '.

  Specifically, when the 18:00 value of the predicted pattern P20 calculated for the current power amount of 80 kWh is 75 kWh, the difference β is 80−75 = 5, and the calculated prediction pattern P20 (18 To 24:00) to correct the continuation pattern from 18:00 (80 kWh).

  Therefore, according to the present embodiment, the target pattern and the predicted pattern obtained by averaging past performance values are corrected to a highly accurate pattern that matches the target value or the current value as a reference, and reliable monitoring can be performed.

(Embodiment 3)
In the first and second embodiments, the target value (target pattern) and the predicted value (predicted pattern) of the daily and monthly power energy usage are obtained, and the power energy is monitored from these target value, predicted value, and current value. However, since the amount of electric energy used (electric energy) is not predicted on an annual basis, the actual state of annual electric energy usage cannot be determined.

  In the present embodiment, a target value in units of electric power in each place, a current power amount, and a predicted value in year units are calculated based on past results, and the actual amount of power energy used in units of years can be grasped.

  The yearly forecast value is corrected according to past to present trends (increase / decrease). In addition, the yearly predicted value can be compared with the monthly target value in a graph format.

(1) System Overview FIG. 14 is a main part configuration diagram of a power energy monitoring system showing an embodiment of the present invention, and a different part from FIG. 1 is that a yearly predicted value calculation unit 70 is added.

  The predicted value calculation unit 70 calculates a predicted value for each year using the current value based on the actual value in a certain sample period. In the prediction, the predicted value is corrected in consideration of the trend (increase / decrease) from the past to the present. The yearly predicted value is obtained as a monthly pattern when the trend graph display unit 50 performs comparison display with the monthly target value. Furthermore, the past actual value, predicted value, and target value are obtained as a monthly unit pattern, and are combined in a year unit and displayed in a graph format.

  The yearly target pattern uses the monthly target value of the previous year obtained in advance by the calculation unit 30 as the target value for the current year. The target value is automatically acquired by the trend graph display unit 50 or set manually on the setting screen.

  In addition, regarding the prediction result in units of years, the alarm reporting unit 60 does not monitor the amount of power consumption in units of years.

(2) Outline of Processing The predicted value calculation unit 70 calculates the average value based on the past actual value (annual report data) of the power energy usage stored in the database 20, and then calculates the monthly value from the past to the present. A tendency (increase / decrease) is examined, and a value corrected according to this tendency is used as a prediction result.

  Among these, as the past actual value of the power energy usage that is the monitoring item, as shown in an example in FIG. 15, actual data (sample data) for the past ten years from the previous year of the prediction year is extracted. However, up to the upper limit of the number of data (defined by the system). As shown in FIG. 16, the target pattern and prediction pattern for each year use a monthly target value, and the prediction pattern calculates the prediction pattern for the current year from the extracted past performance data (sample data). . There are no narrowing conditions. The upper limit on the number of base pattern data is system defined.

  Moreover, the target pattern of the annual energy consumption amount uses the monthly target value of the previous year that is calculated in advance by the calculation unit 30 as the target value for the current year, and this target value is manually set on the setting screen. .

  The yearly forecast value is corrected by examining the monthly trend (increase / decrease) from the past to the present, and performing correction based on the average of past performance data + the trend of the corresponding month (average slope). The presence / absence of this correction is system defined.

  FIG. 17 shows a processing flow of the calculation example of the yearly prediction pattern. First, the actual data for the past 10 years is acquired every month (S21), the average value of the monthly actual data is calculated for each fiscal year, and the average slope (+ for the increasing trend and-for the decreasing trend) is calculated. Further, an increase / decrease tendency is determined (S22, S23). Here, the average slope is obtained by, for example, obtaining slopes of the same month 10 years ago and 9 years ago, slopes of the same month 9 years ago and 8 years ago, and the like until the previous year.

  Next, it is checked whether or not the average slope is increasing (S24). If all the average slopes are increasing, the average slope is added to the previous year value (S25). If all are decreasing, the average slope is subtracted from the previous year's value (S26, S27). Further, when the increasing and decreasing trends are mixed, the average inclination is added to or subtracted from the monthly average (S28). These increase / decrease processes are repeated for 12 months (S29).

  Therefore, the target pattern of the annual energy consumption amount uses the monthly target value of the previous year obtained in advance by the calculation unit 30 as the current year target value, and this target value is manually set on the setting screen. . The forecast pattern of annual energy consumption is calculated based on the past actual value (annual report data) of the electricity energy consumption, and then the monthly trend (increase / decrease) from the past to the present. A value corrected according to this tendency is used as a prediction result.

(3) Specific Pattern Calculation Example FIG. 18 shows an example of annual prediction pattern calculation, and shows a case where the performance data has been on a decreasing trend from the past.

  As shown in the above figure, if there is past performance data P1 to P5 (P1 is one year ago, P5 is five years ago) that has been decreasing from the past, the average slope (average decrease width) for each month is calculated, A pattern obtained by subtracting the average slope from the pattern one year ago is defined as a predicted pattern.

  For example, as a predicted value for December, in FIG. 18, 1 year ago: 60 kW, 2 years ago: 100 kW, 3 years ago: 110 kW, 4 years ago: 130 kW, 5 years ago: 160 kW, and average slope Is ((60-100) + (100-110) + (110-130) + (130-160)) / 4 = -25 kW, the previous December's value is 60 kW, and it is in a decreasing trend. Dec predicted value = 60−25 = 35 kW.

  FIG. 19 shows an example of calculating a prediction pattern in units of years, and shows a case where the performance data has been increasing from the past.

  As shown in the above figure, if there is past performance data P1 to P5 that is increasing from the past (P1 is one year ago, P5 is five years ago), the monthly average slope (average decrease width) is calculated, A pattern obtained by adding an average slope to a pattern one year ago is defined as a predicted pattern.

  For example, as a predicted value for December, in FIG. 19, if 1 year ago: 130 kW, 2 years ago: 110 kW, 3 years ago: 100 kW, 4 years ago: 60 kW, 5 years ago: 30 kW, the average slope is ( (130−110) + (110−100) + (100−60) + (60−30)) / 4 = 25 kW, and the previous December's value is 130 kW and is on an increasing trend. Value = 130 + 25 = 155 kW.

  FIG. 20 shows an example of calculating a prediction pattern in units of years, and shows a case where the actual data has a mixed increase / decrease trend from the past to the present.

As shown in the above figure, if there is past performance data P1 to P5 (P1 is one year ago and P5 is five years ago) in which increasing / decreasing trends are mixed from the past to the present, the average monthly inclination ( (Average decrease width) is calculated, and a pattern obtained by adding (subtracting) the average slope from the average pattern of the past results is set as the predicted pattern.

  For example, as a predicted value for December, in FIG. 20, assuming that 1 year ago: 30 kW, 2 years ago: 60 kW, 3 years ago: 160 kW, 4 years ago: 130 kW, 5 years ago: 110 kW, the average slope is ( (30−60) + (60−160) + (160−130) + (130−110)) / 4 = −20 kW, and since the average value = 98 kW, the predicted value in December of this year = 98−20 = 78 kW It becomes.

(4) Trend graph display example in units of years FIG. 21 shows an example in which the actual value, predicted value, and target value of the power energy usage are displayed in a graph format. In the figure, D1 indicates the actual value for each month, D2 indicates the cumulative total of the actual values, D3 indicates the target value for each month, and D4 indicates the predicted value for each month. In addition, this drain graph display example exemplifies the case where there are two customers (measurement points) such as factories in the actual value D1 of the electric energy consumption, and as a breakdown, the actual value at each customer is D1 _A , It is shown as D1 _B , and this display is appropriately changed according to the number of consumers. Further, the cumulative value D2, the target value D3, and the predicted value D4 are the total values of the two consumers.

  In this way, by comparing and displaying the monthly target value in a graph format, it is possible to easily grasp the long-term (yearly) power energy consumption, manage the target, and raise awareness of energy-saving activities. Can be planned.

10 Electricity and energy consumption collection part 20 Database (DB)
30 day, monthly target value calculation unit 40 day, monthly target value calculation unit 50 trend graph display unit 60 alarm reporting unit 70 year unit predicted value calculation unit

Claims (6)

A power energy monitoring system that supports consumer energy conservation activities by monitoring consumer energy consumption,
A database that sequentially stores the current value and past actual value of customer's power energy usage collected at regular intervals;
A target value calculation unit that obtains a target value (target pattern) of power energy usage in units of days and months based on actual values in a sample period with power energy usage stored in the database;
A predicted value calculation unit for obtaining a predicted value (prediction pattern) of power consumption in daily and monthly units from a current value and a past actual value of power energy usage stored in the database;
From the current value and the past actual value of power energy usage stored in the database, a yearly predicted value calculation unit that obtains a predicted value of yearly power energy usage as a monthly prediction pattern;
A trend graph display unit for displaying the current value and actual value of power energy consumption stored in the database, the target value obtained by the target value calculation unit and the predicted value obtained by the predicted value calculation unit in a trend graph format; ,
An alarm reporting unit that issues an alarm when the predicted value calculated by the predicted value calculating unit deviates from the target value calculated by the target value calculating unit ;
The yearly predicted value calculation unit obtains the past actual value of the power energy consumption stored in the database every month, calculates the average value of the monthly actual value for each year, and The average slope of each month is calculated, and it is determined whether the monthly average slope is increasing or decreasing. If the monthly average slope is increasing, the monthly average slope is added to the same month value of the previous year. If the average slope for each month is decreasing, subtract the average slope for each month from the previous year's same-month value, and if there are both increasing and decreasing trends in the monthly average slope, The average value for each and the average slope for each month are added or subtracted to obtain a predicted value of annual energy consumption,
The trend graph display unit is a graph format in which the past actual pattern in units of months, the prediction pattern in units of months, and the target pattern in units of months acquired automatically or manually are combined in units of years. Electric power energy monitoring system characterized by displaying in .   The daily target pattern calculation by the target value calculation unit and the calculation of the daily prediction pattern by the prediction value calculation unit comprise means for using the actual value combined with the day type. The power energy monitoring system according to 1.   The monthly target pattern calculation by the target value calculation unit and the calculation of the monthly prediction pattern by the prediction value calculation unit comprise means for using the actual values having the same weekly configuration. The power energy monitoring system according to 1.   The power energy according to any one of claims 1 to 3, wherein the target value calculation unit and the predicted value calculation unit include means for performing calculation by narrowing down and averaging the actual values. Monitoring system.   The target value calculation unit includes means for calculating a ratio between the value of the final time of the calculated target pattern and the target value and correcting the deviation from the target value by multiplying the calculated target pattern by this ratio. The power energy monitoring system according to any one of claims 1 to 4, wherein   The predicted value calculating unit includes means for calculating a difference between the calculated predicted pattern and the current value, and correcting a deviation from the current value by adding or subtracting the difference to the calculated predicted pattern. The power energy monitoring system according to any one of claims 1 to 4.

Images