Power demand prediction system and power demand prediction method

Abstract

According to the present invention, building information (104) includes: basic building-information (101) that includes information on the size of a building; maximum annual power that indicates power consumption at a time. within one year. when the power consumption of the building is at a maximum level; and a power consumption amount of the building during a predetermined period within the one year. A processing device (21) extracts, from past power data (103), power data of a similar building for which the basic building-information (101) is similar to that of a target building. The processing device (21) generates power data of the target building by correcting the power data of the similar building so that, for the power data of the similar building and the power data of the target building, there is a match between the maximum annual power consumption amounts and also between the power consumption amounts during the predetermined period.

Description

Electricity demand forecasting system and electricity demand forecasting method

[0001] The present disclosure relates to a power demand forecasting system and a power demand forecasting method for forecasting power data of a target building.

[0002] In order to create an installation plan for a distributed power supply system that is highly effective for the installation of the facility, it is necessary to have time-series data (power data) on the long-term power and heat load of the target building where the facility will be installed. In order to obtain the actual power data of the target building, it is necessary to install measuring instruments and a storage device to store the data. However, it is difficult to install these at the facility design stage before installation, and even if power data exists, it is difficult to obtain it at the early design stage, when the facility is still being sold to the building owner.

[0003] Japanese Patent Laid-Open Publication No. 2009-189085 (Patent Document 1) discloses a system for predicting power and heat loads. This prediction system predicts power and heat loads based on information about the power and heat demand patterns of a target building, the total floor area of ​​the target building, the maximum and minimum loads for power and heat, and the monthly integrated loads for power and heat.

[0004] JP 2009-189085 A

[0005] However, in all of the prediction methods disclosed in Patent Document 1, the prediction of power data is performed with emphasis on the prediction accuracy of only one of the monthly power consumption amount and the annual maximum power consumption. Therefore, the prediction system of Patent Document 1 has a problem in that it is not possible to predict power data with emphasis on the prediction accuracy of both.

[0006] The electricity charges that are taken into consideration when designing a power supply system are determined based on both monthly power consumption and annual maximum power consumption. Therefore, if the prediction accuracy of either is low, the effect of introducing the designed system in reducing electricity charges may be reduced.

[0007] The present disclosure has been made to solve such problems, and the purpose of the present disclosure is to provide an electricity demand forecasting system and an electricity demand forecasting method that can predict electricity data for a target building while improving the prediction accuracy of both the amount of electricity consumed for a specified period and the annual maximum electricity.

[0008] The power demand forecasting system disclosed herein is a system for forecasting power data for a target building. The power demand forecasting system includes a storage device and a processing device. The storage device stores building information for the target building and historical power data recording power data and building information for each of a plurality of buildings different from the target building. The processing device predicts the power data for the target building based on the building information and historical power data for the target building. The power data is time-series data showing trends in power consumption over a year. The building information includes basic building information including information about the size of the building, annual maximum power indicating the power consumption at the time when power consumption in the building was maximum in the year, and the amount of power consumption in the building for a predetermined period of the year. The processing device extracts power data for similar buildings whose basic building information is similar to that of the target building from the historical power data. The processing device corrects the power data of the similar buildings to generate power data for the target building so that the annual maximum power and the amount of power consumption for the predetermined period match between the power data of the similar buildings and the power data of the target building.

[0009] The power demand forecasting method disclosed herein is a method for forecasting power data of a target building. The power demand forecasting method includes the steps of: storing building information of the target building and historical power data recording power data and building information for each of a plurality of buildings different from the target building; and predicting the power data of the target building based on the building information and historical power data of the target building. The power data is time-series data showing trends in power consumption over a year. The building information includes basic building information including information about the size of the building, an annual maximum power indicating the power consumption at the time when the power consumption of the building was at its maximum in the year, and the amount of power consumption of the building for a predetermined period in the year. The forecasting step includes the steps of extracting power data of similar buildings having basic building information similar to that of the target building from the historical power data; and correcting the power data of the similar buildings to generate power data of the target building so that the annual maximum power and the power consumption for the predetermined period match between the power data of the similar buildings and the power data of the target building.

[0010] According to the present disclosure, it is possible to predict power data for a target building while improving the prediction accuracy of both the amount of power consumption for a predetermined period and the annual maximum power.

[0011] FIG. 1 is a functional block diagram of a power demand forecasting system according to a first embodiment; FIG. 2 is a hardware configuration diagram of the power demand forecasting system; FIG. 3 is a flowchart of processing executed by the power demand forecasting system; and FIG. 4 is a graph for explaining a method for predicting power demand. FIG. 4 is a functional block diagram of a power demand forecasting system according to a second embodiment; and FIG. 5 is a flowchart of processing executed by the power demand forecasting system. and FIG. 6 is a graph for explaining overshoot and undershoot of power demand.

[0012] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. While several embodiments will be described below, it was originally intended that the configurations described in each embodiment be combined as appropriate. Note that identical or corresponding parts in the drawings will be designated by the same reference numerals, and their description will not be repeated.

[0013] First Embodiment FIG. 1 is a functional block diagram of a power demand forecasting system 100 according to a first embodiment.

[0014] The power demand forecasting system 100 is a system for forecasting power data for a target building (a building that is the subject of forecasting). Here, the power data (also referred to as "power demand") refers to time-series data (each time of 24 hours x 365 days) that indicates the transition of power consumption in the building over a one-year period.

[0015] The buildings include various types of buildings such as office buildings, commercial buildings, factories, schools, hospitals, etc. The building power data is a one-year trend of the power consumption of various devices and equipment installed in the building, such as air conditioning equipment, lighting, hot water supply equipment, and office equipment.

[0016] The power demand forecasting system 100 includes a basic power demand estimation unit 110, past power data 103, and building information 104. The basic power demand estimation unit 110 of the power demand forecasting system 100 predicts power data of a target building based on the building information 104 and past power data 103 of the target building.

[0017] The building information 104 is information about the target building. The building information 104 includes basic building information 101 and electricity trading information 102. The basic building information 101 includes information about the size of the building. Specifically, the information about the size of the building is the total floor area of ​​the building (hereinafter also simply referred to as "total floor area"), but may also be information indicating other sizes of the building. The basic building information 101 further includes the area in which the building is located (hereinafter also simply referred to as "area") and the use of the building (hereinafter also simply referred to as "use").

[0018] The building basic information 101 correlates with the power consumption (power data) of a building. For example, the larger the total floor area of ​​a building, the greater the power consumption of the building, and the smaller the total floor area of ​​a building, the less the power consumption of the building. The region in which the building is located is, for example, Hokkaido, Tohoku, Kanto, Kinki, etc., but it may also be classified by prefecture, such as Tokyo, Aichi, Osaka, etc. Power consumption trends also differ depending on these regions. For example, Hokkaido has lower temperatures than Kyushu, so power consumption is greater in winter. Uses include the above-mentioned office buildings, commercial buildings, factories, schools, hospitals, etc. Power data trends also differ depending on the use.

[0019] The past power data 103 records power data and building information (total floor area of ​​the building, the area where the building is located, the purpose of the building, etc.) for each of a plurality of buildings different from the target building.

[0020] The power trading information 102 includes the target building's annual maximum power and the target building's power consumption for a predetermined period. The annual maximum power indicates the building's maximum power consumption for one year (the power consumption at the time of day when power consumption is at its maximum in one year). The power consumption for a predetermined period is the power consumption for the building for a predetermined period in one year. In this embodiment, the predetermined period is one month, i.e., each month from January to December.

[0021] That is, for a target building, information exists on the total floor area, region, use, annual maximum power consumption, and monthly power consumption from January to December, but no power data (time-series data showing the trend in power consumption in the building over the course of a year) exists. For this reason, in this embodiment, information such as past power data of other buildings is used to predict the power data of the target building.

[0022] The basic power demand estimation unit 110 includes a past data acquisition unit 111, an annual maximum power adjustment unit 112, and a monthly power amount adjustment unit 113. The past data acquisition unit 111 performs the process of S101 in FIG. 3 , which will be described later. The annual maximum power adjustment unit 112 performs the process of S102. The monthly power amount adjustment unit 113 performs the processes of S103 and S104. This allows the power data of the target building to be predicted. This will be described in more detail later using FIG. 3.

[0023] 2 is a hardware configuration diagram of the power demand forecasting system 100. The power demand forecasting system 100 includes a processor 21 as a processing device, a memory 22, a storage device 23, and a communication device 24. These are connected to each other via a bus 26 so as to be able to communicate with each other.

[0024] The processor is, for example, a CPU (Central Processing Unit). The memory 22 may be configured to include a ROM (Read Only Memory) and a RAM (Random Access Memory). The storage device 23 may be a non-volatile storage device, such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive).

[0025] The processor 21 loads a program stored in the ROM or storage device 23 into the RAM and executes it to realize various functions of the power demand forecasting system 100. The processor 21 performs processes such as predicting power data for a target building. The ROM or storage device 23 stores a program describing the processing procedures of the power demand forecasting system 100. The storage device 23 stores building information 104 and past power data 103.

[0026] The RAM serves as a working area when the processor executes a program, and temporarily stores the program, data used for executing the program, and the like. The power demand forecasting system 100 can be connected to other devices, such as a terminal (not shown), via a communication device 24, either wired or wirelessly. The terminal is, for example, a personal computer or a mobile terminal such as a smartphone. Various information held by the power demand forecasting system 100 can be displayed on the screen of the terminal. It is also possible to issue instructions, such as to execute a program, to the power demand forecasting system 100 from the terminal.

[0027] Furthermore, the power demand forecasting system 100 may include a storage device external to the power demand forecasting system 100. This storage device may be an external storage device such as an HDD, or may be a storage device provided within a server device (such as a file server). The storage device may be configured with multiple storage areas within a single storage device, or may be configured with multiple storage devices. The server device (file server) may also be configured with multiple server devices.

[0028] The process will be described in detail below with reference to a flowchart. Fig. 3 is a flowchart of the process executed by the processor 21 of the power demand forecasting system 100. This process may be executed based on a command from a user via a terminal, for example.

[0029] When this process starts, the processor 21 (past data acquisition unit 111) extracts power data of similar buildings having basic building information similar to that of the target building from the past power data 103 in S101.

[0030] For example, if the building basic information 101 of the target building is a total floor area of ​​"A1" m ２ , area "Tokyo", and use "office building", power data of a building (referred to as a "similar building") having basic building information similar to this basic building information is extracted. For example, if the total floor area is "A1" m ２ Buildings with all the same basic building information, such as "Tokyo" in the area and "office building" in the use field, may be extracted as similar buildings, or buildings with the same area and use field and a similar total floor area (for example, a total floor area within the range of A1 x 90% to A1 x 110%) may be extracted as similar buildings. Alternatively, buildings in areas such as Kanagawa and Chiba, which are close to the area of ​​Tokyo, may also be extracted as similar buildings.

[0031] The processor 21 generates the power data of the target building by correcting the power data of the similar building so that the annual maximum power of the similar building and the power data of the target building match, and so that the amount of power consumed for a predetermined period match. This process is realized by steps S102 to S104 below. Note that in this embodiment, "match" does not necessarily mean that the two are equal, but may also mean that the two values ​​are close (for example, a difference of about 3% may be considered to be a match).

[0032] In S102, the processor 21 (annual maximum power adjustment unit 112) generates corrected power data by multiplying the value obtained by dividing the annual maximum power of the target building by the annual maximum power of a similar building by the data for each time of the power data of the similar building.

[0033] For example, if the target building's annual maximum power consumption is P2 and similar buildings' annual maximum power consumption is PX, each value of the similar building's power data is multiplied by P2 / PX to generate corrected power data. As a result, each value of the similar building's power data is multiplied by (P2 / PX). Furthermore, the annual maximum power consumption of the corrected power data is P2 by multiplying PX by P2 / PX, which can be made to match the target building's annual maximum power consumption.

[0034] In S103, the processor 21 (monthly power amount adjustment unit 113) calculates a coefficient corresponding to each piece of corrected power data, which is weighted so that the power consumption during the peak time period including the time when the annual maximum power of the target building is measured does not change, and is weighted so that the power consumption amount for a predetermined period between the corrected power data and the power data of the target building matches. Here, the predetermined period = January, February, March, ..., December. In S104, the processor 21 (monthly power amount adjustment unit 113) multiplies each piece of corrected power data by the coefficient calculated for each piece of corrected power data to generate power data for the target building.

[0035] A specific explanation will be given below using FIG. 4. FIG. 4 is a graph for explaining a method for predicting power demand (power data). The diagram at the bottom left of FIG. 4 shows an example of data for one day among the corrected power data generated from the power data of similar buildings. In this example, data for a day including the annual maximum power P2 is shown. The minimum power on this day is assumed to be P1. Furthermore, in this example, power data for the target building for August, when the annual maximum power P2 was detected, is generated.

[0036] The diagram on the upper left shows the coefficient by which the corrected power data is multiplied. When the coefficient = 1, the corrected power data does not change. In this example, the annual maximum power P2 is detected at 12:00. The peak time period, which includes the time (12:00) when the annual maximum power of the target building was measured, is set to 10:00 to 14:00. In this example, a four-hour period is set as the peak time period, but this is not limited to this and any length of time (such as three hours) may be used.

[0037] In this case, 10:00 and 14:00 are set as fixed points with a coefficient of 1. Furthermore, within the range from midnight to midnight, fluctuating points are set between midnight and 10:00 (fixed point), and between 14:00 (fixed point) and midnight. In this example, fluctuating points are set at 4:00 and 20:00.

[0038] Next, the coefficient K1 is determined so that the power consumption of the target building in August matches the power consumption of the corrected power data for August. The coefficient K1 is the coefficient for the fluctuation points of 4:00 and 20:00.

[0039] In this case, the coefficients for midnight to 4 o'clock and 8 o'clock to midnight are also set to coefficient K1. Also, for example, the coefficients for 4 o'clock to 10 o'clock may be determined by linear interpolation based on the values ​​of the coefficient for 4 o'clock (= K1) and the coefficient for 10 o'clock (= 1). The coefficients for 2 o'clock to 8 o'clock may be determined by linear interpolation based on the values ​​of the coefficient for 2 o'clock (= 1) and the coefficient for 8 o'clock (= K1).

[0040] For example, as shown in the diagram on the upper right, the coefficients for August (coefficients for each time from midnight to midnight) are determined. The processor 21 multiplies each of the corrected power data (diagram on the lower left) by each of the coefficients for August (diagram on the upper right) to generate power data (for August) for the target building as shown in the diagram on the lower right.

[0041] The coefficient data in the upper right diagram is multiplied by the corrected power data for each day from August 1st to 31st. This generates the power data for the target building from August 1st to 31st. At this time, K1 is calculated so that the power consumption of the generated power data from August 1st to 31st matches the power consumption of the target building in August.

[0042] The peak hours are from 10:00 to 14:00, and the coefficient is 1. Therefore, the power consumption between 10:00 and 14:00 matches the power consumption between the bottom left and bottom right diagrams. As a result, the annual maximum power consumption P2 at 12:00 also remains unchanged. On the other hand, the power consumption between midnight and 4:00 and between 20:00 and 24:00 has changed by a factor of K1.

[0043] The power data is calculated in the same way for months other than August. Note that the calculation example described above is merely an example. The coefficients may be calculated by any method as long as the weighted coefficients are calculated so that the power consumption during peak hours does not change and the power consumption amounts for each month are consistent. For example, the coefficients for each month may be calculated by using the difference in the power consumption amounts for each month (e.g., August) after multiplying the above-mentioned correction data by each coefficient as the objective function and solving a linear programming problem with each coefficient as an operating variable. In this way, by weighting (calculating coefficients) so as not to change the power demand (power data) during peak hours, the monthly power consumption can be adjusted without changing the annual maximum power.

[0044] As described above, in the first embodiment, the power demand forecasting system 100 is a system that forecasts power data for a target building. The power demand forecasting system 100 includes a storage device 23 and a processor 21 as a processing device. The storage device 23 stores building information 104 for the target building and historical power data 103 that records power data and building information for each of a plurality of buildings different from the target building. The processor 21 forecasts the power data for the target building based on the building information 104 and the historical power data 103 for the target building. The power data is time-series data indicating trends in power consumption over a one-year period. The building information 104 includes basic building information 101 that includes information about the size of the building, annual maximum power indicating the power consumption at the time when the building's power consumption was at its maximum in a year, and the amount of power consumption for the building for a predetermined period in a year. The processor 21 extracts power data for similar buildings whose basic building information 101 is similar to that of the target building from the historical power data 103. The processor 21 corrects the power data of the similar building to generate power data of the target building so that the annual maximum power and the amount of power consumption for a specified period match between the power data of the similar building and the power data of the target building. The information regarding the size of the building is the total floor area of ​​the building. The building basic information 101 further includes the area in which the building is located and the purpose of the building. The specified period is one month.

[0045] The electricity charges taken into consideration when planning distributed power generation equipment (power generation equipment, power storage equipment) for a power supply system are determined based on both monthly power consumption and annual maximum power consumption. Therefore, if either of these predictions deviates from the forecast, the electricity bill reduction effect of the designed system may be reduced. In order to create an installation plan for distributed power generation equipment that is highly effective for implementation, it is necessary to improve the accuracy of both of these predictions. By adjusting both the power consumption for a specified period (e.g., monthly) and the annual maximum power consumption as in this embodiment, it is possible to predict the power data (24-hour x 365-day power demand) of the target building while improving the prediction accuracy of both the power consumption for the specified period and the annual maximum power consumption. This improves the electricity bill reduction effect of the designed power supply system.

[0046] The processor 21 generates corrected power data by multiplying the power data of the similar buildings at each time by a value obtained by dividing the annual maximum power of the target building by the annual maximum power of the similar building. The processor 21 calculates a coefficient corresponding to each corrected power data, which is weighted so that the power consumption during peak hours including the time when the target building's annual maximum power is measured does not change, and so that the corrected power data and the power data of the target building match the amount of power consumption over a predetermined period. The processor 21 generates power data for the target building by multiplying each corrected power data by the coefficient calculated for each corrected power data. In this way, by calculating a weighted coefficient that does not change the power data during peak hours, the amount of power consumption over a predetermined period (monthly power consumption) can be adjusted without changing the annual maximum power. This makes it possible to predict the power data of the target building while improving the prediction accuracy of both the amount of power consumption over a predetermined period and the annual maximum power.

[0047] 5 is a functional block diagram of a power demand forecasting system 100a according to a second embodiment. The power demand forecasting system 100a according to the second embodiment further includes a power demand fluctuation data aggregation unit 121, a derived power demand estimation unit 122, and power demand fluctuation data 131.

[0048] The power demand fluctuation data aggregation unit 121 performs the process of S201 in Fig. 6 (described later) to generate power demand fluctuation data 131. The derived power demand estimation unit 122 performs the processes of S202 and S203. This generates multiple derived scenarios for the power data of the target building. The following description will be given with reference to Fig. 6.

[0049] 6 is a flowchart of the process executed by the power demand forecasting system 100 a. This flowchart shows the process executed by the power demand fluctuation data aggregation unit 121 and the derived power demand estimation unit 122. The process executed by the basic power demand estimation unit 110 is the same as the process described using FIG. 3.

[0050] In S201, the processor 21 (electricity demand fluctuation data aggregation unit 121) calculates estimated data that estimates the relationship between the building basic information 101 and the average value and variance of time-series data showing the trend in daily power consumption based on the past power data 103, and saves this as electricity demand fluctuation data 131. The estimated data is calculated for each month.

[0051] For example, it is assumed that the historical power data 103 contains a plurality of buildings (buildings X1 to X3) with total floor area = A, area = Tokyo, and use = office building as the building basic information 101.

[0052] To calculate estimated data for August, a total of 93 days' worth of time series data (31 days x 3) is obtained: 31 days' worth of time series data from August 1st to 31st for building X1, 31 days' worth of time series data from August 1st to 31st for building X2, and 31 days' worth of time series data from August 1st to 31st for building X3. The average and variance for each time (0:00 to 24:00) for this total of 93 days' worth of time series data are calculated.

[0053] This is stored as estimated data indicating the relationship between the building basic information (total floor area = A, region = Tokyo, use = office building) and the average value and variance for August in the electricity demand fluctuation data 131. Since variance is calculated for each hour, for example, if there is a large variance in the data around 12 o'clock, the variance around 12 o'clock will also be large.

[0054] In this way, the relationship between the total floor area, region, and use of a building and the average value and variance of the power data (by month) can be derived for each month from the past power data 103. The relationship between the total floor area, region, and use of a building and the average value and variance of the power data can be statistically determined and stored as power demand fluctuation data 131 in the form of a table, regression equation, or the like.

[0055] In S202, the processor 21 (derived power demand estimation unit 122) calculates target estimated data by estimating the average value and variance of time-series data showing the transition of daily power consumption of the target building based on the estimated data (power demand fluctuation data 131) and the power data of the target building. The target estimated data is calculated for each month.

[0056] For example, if the basic building information of the target building is total floor area = A, region = Tokyo, and use = office building, the average value and variance corresponding to the data total floor area = A, region = Tokyo, and use = office building stored in the electricity demand fluctuation data 131 are set as the target estimated data for the target building.

[0057] The processor 21 corrects the power data of the target building based on the power data of the target building and the target estimation data. Specifically, in S203, the processor 21 (derived power demand estimator 122) estimates a trend in the maximum value of power consumption (referred to as the "maximum scenario"), a trend in the minimum value of power consumption (referred to as the "minimum scenario"), and a trend in the average value of power consumption (referred to as the "average scenario") based on the power data of the target building and the target estimation data (average value and variance). This will be explained in detail below using FIG. 7.

[0058] 7 is a graph illustrating fluctuations in power demand. As described above, the basic power demand estimation unit 110 calculates power data (power demand) for the target building. Then, by performing the processes in S201 to S203 described above, the average value and variance of the daily power data (target estimated data) for the target building for each month from January to December are calculated. Then, by applying the target estimated data for each month to the power data for the target building, multiple scenarios (maximum scenario, minimum scenario, average scenario) are obtained.

[0059] For example, when the target estimated data for August (mean value and variance) is applied to the power data for the target building on August 15th, the multiple scenarios obtained are as shown in the graph in Figure 7. The maximum scenario for the target building on August 15th is graph L2, the minimum scenario is graph L3, and the average scenario is graph L1. At 12 o'clock, the power consumption of the maximum scenario is P2, the power consumption of the minimum scenario is P3, and the power consumption of the average scenario is P1. Here, P2 > P1 > P3.

[0060] The maximum scenario (the trend in maximum power consumption) is a scenario in which the average scenario exceeds the average. The minimum scenario (the trend in minimum power consumption) is a scenario in which the average scenario (the trend in average power consumption) exceeds the average.

[0061] For example, each scenario may be generated based on the relationship between the mean value μ and the standard deviation σ in a normal distribution. For example, the scenarios may be generated so that the maximum scenario is μ+σ (upside), the average scenario is μ, and the minimum scenario is μ-σ (downside), or the maximum scenario is μ+2σ, the minimum scenario is μ-2σ, or the maximum scenario is μ+3σ, and the minimum scenario is μ-3σ.

[0062] The maximum scenario may be any scenario that causes the power data of the target building to deviate upward based on the variance. The minimum scenario may be any scenario that causes the power data of the target building to deviate downward based on the variance. The average scenario may correct the power data of the target building using an average value so that the power data of the target building becomes standard power data that can be estimated from the basic building information of the target property, or the power data of the target building may be used as the average scenario as is.

[0063] By deriving scenarios in this way, it is possible to obtain a worst-case scenario (maximum scenario) in which the electricity bill for the target building will be the highest, a best-case scenario (minimum scenario) in which the electricity bill for the target building will be the lowest, and a standard-case scenario (average scenario).By estimating the worst-case or best-case scenario, when considering the introduction of new power generation equipment such as solar power generation and energy storage equipment, it is possible to plan the equipment while taking uncertainty into account.

[0064] As described above, in the second embodiment, the processor 21 calculates estimated data that estimates the relationship between the building basic information 101 and the average value and variance of time-series data showing the trend in daily power consumption based on past power data. The processor 21 calculates target estimated data that estimates the average value and variance of time-series data showing the trend in daily power consumption of the target building based on the estimated data and the power data of the target building. The processor 21 corrects the power data of the target building based on the power data of the target building and the target estimated data. The processor 21 estimates the trend in maximum power consumption, the trend in minimum power consumption, and the trend in average power consumption based on the power data of the target building and the target estimated data. The processor 21 calculates the target estimated data for each month.

[0065] In this way, publicly available past power data from other buildings is compiled to determine the uncertainty of cases where power demand (power data) is higher or lower than average. The uncertainty of the power data corresponding to the target building is then used to change the predicted power data. This makes it possible to make predictions that take into account not only average power data but also the uncertainty of power data that exists during operation, such as power data that fluctuates above or below the average, enabling designs that are highly effective in reducing costs throughout the entire lifecycle.

[0066] The embodiments disclosed herein are intended to be combined as appropriate within the scope of any technical inconsistency. The embodiments disclosed herein should be considered to be illustrative and not restrictive in all respects. The technical scope of the present disclosure is defined by the claims, not the description of the above-mentioned embodiments, and is intended to include all modifications within the meaning and scope of the claims.

[0067] 21 Processor, 22 Memory, 23 Storage device, 24 Communication device, 26 Bus, 100, 100a Power demand forecasting system, 101 Building basic information, 102 Power buying and selling information, 103 Past power data, 110 Basic power demand estimation unit, 111 Past data acquisition unit, 112 Annual maximum power adjustment unit, 113 Monthly power amount adjustment unit, 121 Power demand fluctuation data aggregation unit, 122 Derived power demand estimation unit, 131 Power demand fluctuation data.

Claims

1. A power demand forecasting system that predicts power data for a target building, comprising: a storage device that stores building information for the target building and past power data that records the power data and the building information for each of a plurality of buildings different from the target building; a processing device that predicts the power data of the target building based on the building information and the past power data of the target building, wherein the power data is time series data that shows trends in power consumption over a year, and the building information includes basic building information including information about the size of the building, an annual maximum power that shows the maximum power consumption of the building over a year, and the amount of power consumption of the building for a specified period within the year, and the processing device: extracts from the past power data the power data of similar buildings whose basic building information is similar to that of the target building, and corrects the power data of the similar buildings to generate the power data of the target building so that the annual maximum power and the power consumption for the specified period match between the power data of the similar buildings and the power data of the target building.

2. The power demand forecasting system described in claim 1, wherein the processing device: generates corrected power data by multiplying the data for each time in the power data of the similar building by a value obtained by dividing the annual maximum power of the target building by the annual maximum power of the similar building; calculates a coefficient corresponding to each of the corrected power data, which is weighted so that the power consumption during the peak time period including the time when the annual maximum power of the target building does not change and is weighted so that the amount of power consumed during the specified period between the corrected power data and the power data of the target building matches; and generates the power data of the target building by multiplying each of the corrected power data by the coefficient calculated corresponding to each of the corrected power data.

3. The processing device calculates estimated data that estimates the relationship between the building basic information and the average value and variance of time series data showing the trend in daily power consumption based on the past power data, calculates target estimated data that estimates the average value and variance of time series data showing the trend in daily power consumption of the target building based on the estimated data and the power data of the target building, and corrects the power data of the target building based on the power data of the target building and the target estimated data. The power demand forecasting system described in claim 1.

4. The electricity demand forecasting system of claim 3, wherein the processing device estimates trends in maximum electricity consumption, minimum electricity consumption, and average electricity consumption based on the electricity data of the target building and the target estimation data.

5. The electricity demand forecasting system according to claim 3 or claim 4, wherein the processing device calculates the target estimation data on a monthly basis.

6. An electricity demand forecasting system as described in any one of claims 1 to 5, wherein the information regarding the size of the building is the total floor area of ​​the building, the basic building information further includes the area in which the building is located and the use of the building, and the specified period is one month.

7. A power demand forecasting method for predicting power data for a target building, comprising the steps of: storing building information for the target building and past power data recording the power data and the building information for each of a plurality of buildings different from the target building; and predicting the power data of the target building based on the building information and the past power data of the target building, wherein the power data is time series data showing trends in power consumption over a year, and the building information includes basic building information including information about the area of ​​the building, an annual maximum power showing the power consumption at the time when power consumption in the building was at its maximum in the year, and the amount of power consumption in the building for a specified period in the year, and the predicting step comprises the steps of: extracting from the past power data the power data of a similar building whose basic building information is similar to that of the target building; and correcting the power data of the similar building to generate the power data of the target building so that the annual maximum power and the power consumption for the specified period match between the power data of the similar building and the power data of the target building.