JP7489904B2 - Regional energy management system, regional energy management device, facility energy management device, demand forecasting method, and demand forecasting program - Google Patents

Description

This disclosure relates to a local energy management system that manages the supply and demand of electricity in a region, a local energy management device, a facility energy management device, a demand forecasting method, and a demand forecasting program.

Technologies that aim to make efficient use of electricity throughout an entire region (community), such as smart communities, are being introduced and examined. In such regions, electricity is shared within the region to make effective use of the local power generation and storage facilities.

Generally, in such areas, a local energy management device is installed, which manages the supply and demand of electricity in the area. The local energy management device predicts the demand and supply of electricity in the area and formulates a supply and demand plan in order to take measures such as requesting consumers in the area to reduce their power consumption or purchasing electricity from a power company or the electricity market if there is a power shortage even after power exchange. To exchange electricity effectively within the area, it is desirable to accurately predict electricity demand.

Patent Document 1 discloses a technology for improving the accuracy of power demand prediction. The technology described in Patent Document 1 predicts power demand and pedestrian flow using past actual values. Next, the technology described in Patent Document 1 obtains the difference between the measured power demand value obtained every hour and the predicted power demand value as the power demand prediction error, and estimates the pedestrian flow based on the position information of mobile devices, demographic values, traffic volume, and the number of passengers boarding and alighting at each station, which are obtained every hour, and calculates the pedestrian flow prediction error, which is the difference between the predicted pedestrian flow value and the estimated pedestrian flow value. Then, in the technology of Patent Document 1, if it is determined that there is a correlation between the power demand prediction error and the pedestrian flow prediction error, the pedestrian flow prediction error is used to correct the power demand prediction value.

JP 2017-208952 A

In recent years, teleworking and business voluntary restraints have been implemented as a measure against infectious diseases, and such social conditions are expected to continue in the future. Teleworking and business voluntary restraints affect the flow of people, which in turn affects electricity demand. According to the technology described in the above-mentioned Patent Document 1, it is possible to calculate electricity demand by taking into account fluctuations in the flow of people, but the location information of mobile devices, demographic values, traffic volume, and the number of passengers boarding and alighting at each station, which are the basis for the estimated value of the flow of people, do not necessarily affect electricity demand immediately. For example, the time period when the number of passengers boarding and alighting at the nearest station to an office building or factory is high is different from the time period when the power demand of the office building or factory is at its peak. For this reason, with regard to the number of passengers boarding and alighting at the nearest station, the time difference between the time when the power demand is predicted and the time when the number of passengers boarding and alighting at the nearest station is counted may affect the prediction error of the power demand. On the other hand, for example, since the demographic values are not frequently updated, the time difference between the time when the power demand is predicted and the time when the demographic values are calculated is not a big problem. In Patent Document 1, since the above-mentioned time difference is not taken into account, the accuracy of the prediction of the power demand may be insufficient.

The present disclosure has been made in consideration of the above, and aims to provide a regional energy management system that can improve the accuracy of electricity demand forecasts.

In order to solve the above-mentioned problems and achieve the object, the local energy management system according to the present disclosure includes a local energy management device that manages the supply and demand of electricity in an entire area where electricity is exchanged, and a plurality of facility energy management devices that manage electricity in respective areas defined in the area. The facility energy management devices obtain forecast values of electricity demand in their respective corresponding areas and transmit the forecast values to the local energy management device, the local energy management device corrects the forecast values from the facility energy management devices using demand impact information and manages the supply and demand of the entire area using the corrected forecast values, the demand impact information including the number of passengers getting on and off at stations, the local energy management device calculates a correlation value between the demand impact information and the electricity demand for each time difference between a reference time and the time of the electricity demand using the acquired demand impact information and the electricity demand, and sets a time difference used for correcting the forecast value using the calculated correlation value .

The regional energy management system disclosed herein has the effect of improving the accuracy of electricity demand forecasts.

FIG. 1 is a diagram showing a configuration example of a local energy management system according to an embodiment; FIG. 1 is a diagram showing an example of the configuration of a computer system that realizes an AEMS according to an embodiment. Schematic diagram showing changes in electricity demand in offices Schematic diagram showing changes in electricity demand at retail stores Schematic diagram showing changes in electricity demand in a hotel A diagram showing an example of the impact of social conditions on electricity demand due to infection control measures for each business type FIG. 13 is a diagram showing business hours information as an example of first information; FIG. 13 is a diagram showing an example of attendance status of the first information; A diagram showing the relationship between surveillance camera images and predictions of the number of people present on a floor of a building. FIG. 10 is a diagram showing an example of correction of power demand based on the number of passersby recognized from the image of the surveillance camera shown in FIG. FIG. 13 is a diagram showing an example of a relationship between the number of passengers getting on and off at a station and power demand, which is an example of second information; A schematic diagram showing correlation values for each time difference of information FIG. 13 is a diagram showing an example of the second information, which is meteorological information, and its influence on power demand; 11 is a flowchart showing an example of a procedure for generating correction information. A flowchart showing an example of a procedure for correcting power demand in the AEMS according to the embodiment. A schematic diagram showing an example of building and station locations

The following provides a detailed explanation of the local energy management system, local energy management device, facility energy management device, demand forecasting method, and demand forecasting program according to the embodiments, with reference to the accompanying drawings.

Figure 1 is a diagram showing an example of the configuration of a local energy management system according to an embodiment. The local energy management system of this embodiment includes an area energy management system (AEMS) 1, which is a local energy management device, and building energy management systems (BEMS) 2-1 to 2-3, which are facility energy management devices that manage the energy of buildings. The AEMS 1 and each of the BEMS 2-1 to 2-3 are connected via a communication network.

The AEMS 1 manages the supply and demand of electricity for the entire region in which electricity is exchanged within the region. The BEMS 2-1 to 2-3 manage the supply and demand of electricity for the buildings corresponding to each of them. The buildings are examples of areas defined in the region. In the example shown in FIG. 1, there are three buildings in the region, which are three areas, and three BEMSs, BEMS 2-1 to 2-3, corresponding to each building are provided, but the number of buildings in the region that exchange electricity is not limited to three. In the following, an example is shown in which there are multiple buildings as multiple areas within the region, but this is not limiting. Instead of building facilities, there may be areas in the region where multiple factory facilities, large commercial facilities, etc. are provided with corresponding facility energy management devices (xEMS: x Energy Management System), and buildings and these facilities may be mixed within the region.

When managing the energy of an entire region that exchanges power within the region, the demand for power fluctuates depending on the business situation, the attendance status of employees, and the like, particularly in office buildings, factories, large commercial facilities, and the like. Conventionally, power demand has been predicted based on past actual values, but predictions based on past actual values do not provide sufficient accuracy, particularly in situations where teleworking and business voluntary restraints are implemented as infection control measures, or weather warnings are issued. In this embodiment, each BEMS 2-1 to 2-3 predicts power demand based on actual values, and the AEMS 1 corrects the predicted values received from each BEMS 2-1 to 2-3 using various information. Here, depending on the information used for correction, the effect of the value indicated by the information does not necessarily appear on the power demand at the same time as the time corresponding to the value, and there may be a time lag before the effect on power demand appears. In this embodiment, the AEMS 1 corrects the predicted value taking this time lag into account, thereby improving the accuracy of the power demand prediction.

As shown in FIG. 1, the BEMS 2-1 includes a communication unit 21, a demand forecasting unit 22, and a memory unit 23. The communication unit 21 communicates with other devices. The communication unit 21 can communicate with the AEMS 1, for example, via a communication network, as well as with the building security management system 3 and the work management system 4.

The building security management system 3 is a system that manages security in the building managed by the BEMS 2-1, and performs the following: acquiring images from surveillance cameras installed in the building, managing entrances to the building gates, managing entrances and exits to each room in the building, managing elevators, and grasping people's movements using various sensors. The work management system 4 is a work management system for tenants who reside in the building managed by the BEMS 2-1, and manages the attendance status of employees. The work management system 4 is basically provided for each tenant, but all tenants in the building may be managed collectively. In addition, some tenants may not be managed by the work management system 4, and in this case, the BEMS 2-1 may obtain the attendance status of the tenant in which the work management system 4 is provided. The building security management system 3 and the work management system 4 are examples of devices that provide first information. The first information is information that may affect the power in the building managed by the BEMS 2-1, and is information that can be obtained by the BEMS 2-1 and is obtained in the building. In other words, the first information is information that the BEMS 2-1 obtains in the building. Therefore, the first information is also called local information. The first information includes, for example, at least one of the following: images captured by a surveillance camera, information indicating the working patterns of employees of tenants in the building, information indicating business hours of tenants, entry and exit management information for each room in the building, the number of times the building's doors are opened and closed, the number of times the building's windows are opened and closed, the number of times the elevator goes up and down, the number of people entering the building, and the amount and direction of people in each part of the building. The first information is not limited to these, and may be any information that may affect the power of BEMS2-1 and that can be acquired by BEMS2-1.

The communication unit 21 temporarily stores the first information acquired from the building security management system 3 and the work management system 4 in the memory unit 23. The communication unit 21 also reads out the first information stored in the memory unit 23 and transmits it to the AEMS 1, for example, periodically. The device from which the first information is acquired may be either the building security management system 3 or the work management system 4, or may be another device other than the building security management system 3 and the work management system 4.

The communication unit 21 also stores various actual values, including the actual value of the power demand, in the storage unit 23 as actual information. Specifically, the communication unit 21 acquires the measured value of the amount of power usage in the building from a measuring instrument (not shown) and stores it in the storage unit 23 as an actual value of the demand amount together with information indicating the date and time. It is preferable that this measured value (actual value) is stored for each tenant. The communication unit 21 also acquires information such as temperature and humidity from a weather information providing system, which is an example of the external system 5, via the communication network, and stores the acquired information in the storage unit 23 as an actual value of the weather information together with information indicating the date and time. In this way, the actual value stored in the storage unit 23 includes the actual value of the demand amount and the actual value of the weather information. The actual value may also include calendar information indicating the day of the week for each date, whether it is a weekday or a holiday, and the like. In this way, the actual information includes the actual value of the demand amount and the actual value of the weather information. The actual information may also include an actual value of the calendar information. The weather information and calendar information are information used for power prediction in the BEMS 2-1, and the actual information may also include information generally used for predicting power demand in addition to these. Furthermore, the method of acquiring each performance value is not limited to the method acquired by the communication unit 21, but may be input by an operator using an input means of the BEMS 2-1 (not shown).

The communication unit 21 also transmits performance information stored in the memory unit 23 to the AEMS 1. This performance information is used for the analysis in the AEMS 1, which will be described later.

The demand forecasting unit 22 predicts the power demand in the building using the actual information stored in the memory unit 23. That is, the demand forecasting unit 22 obtains a predicted value of the power demand of the corresponding building using the actual value of the power demand in the past. Here, the prediction period for predicting the power demand is one day, and the prediction is performed in units of one day divided into 48 slots of 30 minutes. The prediction period and the unit of prediction are not limited to this. The method of predicting the power demand in the demand forecasting unit 22 may be any method as long as it is a method of predicting based on the actual value of the demand amount. For example, the demand forecasting unit 22 obtains forecast values of temperature, humidity, etc. for the prediction period from the weather information providing system via the communication unit 21, and among the actual values of the power demand stored in the memory unit 23, the actual value of the power demand for one day on a date that is the same season or month as the prediction period, has the same day of the week, and has the highest temperature closest to the forecast value is set as the predicted value for the prediction period. The demand forecasting unit 22 transmits the predicted value of the power demand to the AEMS1 via the communication unit 21.

In the following explanation, the demand forecasting unit 22 forecasts the power demand based on actual information for dates and times when there are no changes in social conditions, such as infectious disease countermeasures, and no special weather warnings have been issued. Instead of excluding days when weather warnings have been issued, the influence of weather warnings can be eliminated by calculating the average power demand for many days in the same season with little difference in temperature, humidity, etc., and on the same day of the week, instead of excluding days on which weather warnings have been issued.

BEMS2-2, 2-3 have the same configuration as BEMS2-1 and perform the same operation as BEMS2-1. Like BEMS2-1, BEMS2-2, 2-3 acquire first information from the building security management system 3 that manages the corresponding building, the tenant work management system 4 that inputs the information to the corresponding building, and the like. However, conditions may differ depending on the building, for example, the building managed by BEMS2-1 may have an entrance gate, while the building managed by BEMS2-2 may not have an entrance gate. For this reason, the contents of the first information may differ for each of BEMS2-1 to 2-3, for example, the first information collected by BEMS2-1 includes the number of people passing through the entrance gate, while the first information collected by BEMS2-2 does not include the number of people passing through the entrance gate.

The AEMS 1 includes a supply and demand management unit 11, an analysis unit 12, a correction unit 13, a communication unit 14, a memory unit 15, and a display unit 16. The communication unit 14 communicates with other devices. For example, the communication unit 14 receives the forecast value, the first information, and the actual information of the power demand corresponding to each of the BEMS 2-1 to 2-3, and stores them in the memory unit 15. The communication unit 14 also stores the second information acquired from the external system 5 outside the regional energy management system in the memory unit 15. The second information is, for example, information acquired from the external system 5, and is information other than the information acquired in the buildings managed by the BEMS 2-1 to 2-3. The second information is, for example, mobile phone location information, the number of passengers boarding and alighting at stations, weather warnings, transportation schedules, information indicating requests from the government to refrain from business and to refrain from going out (government information), public information from corporate public relations (corporate information), etc. The external system 5 is, for example, a weather information providing device, a device providing mobile phone location information, a device providing the number of passengers boarding and alighting at stations, etc. The second information is more general information rather than specific to buildings in the area. For this reason, the second information is also called wide-area information.

The analysis unit 12 analyzes the relationship between the first information and the second information and the power demand using the first information, the second information, and the actual information stored in the memory unit 15, and generates correction information for correcting the power demand using at least one of the first information and the second information using the analysis result. The first information and the second information are both information that may affect the demand for power, and when they are not distinguished from each other, they are referred to as demand influence information. The correction information generated by the analysis unit 12 may indicate a relational expression for calculating a correction value of the power demand using the demand influence information, or may indicate the demand influence information and the correction value of the power demand as a table. In addition, the analysis unit 12 may generate a trained model for calculating a correction value of the power demand from the demand influence information by performing machine learning using the demand influence information and the actual information, and this trained model may be used as the correction information. Note that, although an example in which the AEMS 1 includes the analysis unit 12 has been shown here, the AEMS 1 may not include the analysis unit 12, and a table or a calculation formula for performing the correction may be input by an operator or transmitted from an external device as correction information and stored in the memory unit 15. When a trained model is generated by machine learning using the demand impact information and the actual value of the power demand, for example, the correction unit 13 infers a correction value to be used for correcting the predicted value using the trained model, that is, the correction model, and performs the correction using the inferred correction value. That is, the corrected predicted value is generated by inference using the correction model trained by machine learning using the demand impact information and the actual value. Note that in this case as well, a trained model is generated in which the demand impact information and the actual value of the power demand are input as time-series data so that the correction model takes into account the time difference between the time when the impact of the value indicated by the information appears in the power demand at the same time as the time corresponding to the value, thereby reflecting the time difference in the correction.

The correction unit 13 uses the demand impact information and the correction information to correct the forecast values transmitted from the BEMS 2-1 to 2-3 and stored in the memory unit 15. The correction unit 13 updates the forecast values stored in the memory unit 15 with the corrected values. Note that when the forecast values before correction are used to generate correction information by the analysis unit 12, the forecast values before correction are also left in the memory unit 15 as past forecast values. When the correction unit 13 receives demand impact information corresponding to the demand value at the date and time when correction has already been performed, it corrects the forecast value using the received demand impact information and updates the forecast value stored in the memory unit 15 with the corrected value. In this way, when new demand impact information is received, correction is performed again, so that a correction that reflects the new demand impact information can be performed.

The supply and demand management unit 11 calculates the demand for the entire region using the corrected predicted values of the BEMS 2-1 to 2-3, and creates a supply and demand plan for the region using the calculated demand. That is, the supply and demand management unit 11 manages the supply and demand for the entire region using the corrected predicted values. The display unit 16 displays the supply and demand plan, the predicted value of the demand, and the like. The supply and demand management unit 11 creates a supply and demand plan for the next day, for example, and then sequentially updates the created supply and demand plan using the latest information. For example, each BEMS 2-1 to 2-3 generates a predicted value of the power demand for the next day and transmits it to the AEMS 1, the correction unit 13 of the AEMS 1 corrects these predicted values, and the supply and demand management unit 11 creates a supply and demand plan based on the corrected predicted values and predicted values such as the amount of power generation in the region. After that, when the power demand is corrected again using the latest information for the period corresponding to the supply and demand plan and the predicted value is updated, the supply and demand management unit 11 updates the supply and demand plan using the updated power demand. The supply and demand management unit 11 may also formulate a supply and demand plan for a longer period than the next day, such as a one-week supply and demand plan. Here, an example is shown in which the AEMS 1 displays the supply and demand plan and the predicted value of the demand amount, but these may be transmitted to another device and displayed on the other device.

Here, we will explain an example of the hardware configuration of AEMS1. Figure 2 is a diagram showing an example of the configuration of a computer system that realizes AEMS1 of this embodiment.

As shown in FIG. 2, this computer system includes a control unit 101, an input unit 102, a storage unit 103, a display unit 104, a communication unit 105, and an output unit 106, which are connected via a system bus 107. In FIG. 2, the control unit 101 is, for example, a CPU (Central Processing Unit) or the like, and executes a program in which the processing in each server of this embodiment is described. The input unit 102 is, for example, composed of a keyboard, a mouse, and the like, and is used by the user of the computer system to input various information. The storage unit 103 includes various memories such as a RAM (Random Access Memory) and a ROM (Read Only Memory) and a storage device such as a hard disk, and stores the program to be executed by the control unit 101, necessary data obtained in the process of processing, and the like. The storage unit 103 is also used as a temporary storage area for the program. The display unit 104 is composed of an LCD (Liquid Crystal Display Panel) or the like, and displays various screens to the user of the computer system. The communication unit 105 is a receiver and a transmitter that perform communication processing. The output unit 106 is a printer or the like.

Here, an example of the operation of the computer system until the program that realizes the AEMS1 of this embodiment is in an executable state will be described. In the computer system having the above-mentioned configuration, for example, a program is installed in the storage unit 103 from a CD-ROM or DVD-ROM set in a CD (Compact Disc)-ROM drive or DVD (Digital Versatile Disc)-ROM drive (not shown). Then, when the program is executed, the program read from the storage unit 103 is stored in the storage unit 103. In this state, the control unit 101 executes processing as the AEMS1 of this embodiment according to the program stored in the storage unit 103.

In the above explanation, the program describing the processing is provided on a CD-ROM or DVD-ROM as a recording medium. However, this is not limiting, and depending on the configuration of the computer system and the capacity of the program provided, for example, a program provided via a transmission medium such as the Internet via the communication unit 105 may be used.

The communication unit 14 shown in FIG. 1 is realized, for example, by the communication unit 105 shown in FIG. 2. The supply and demand management unit 11, analysis unit 12, and correction unit 13 shown in FIG. 1 are realized by the control unit 101 executing a program. The memory unit 15 shown in FIG. 1 is realized by the memory unit 103 shown in FIG. 2. The display unit 16 shown in FIG. 1 is realized by the display unit 104 shown in FIG. 2. Note that FIG. 2 is just an example, and the configuration of the computer system is not limited to the example shown in FIG. 2. For example, the computer system does not need to be provided with the output unit 106.

The demand forecasting program of this embodiment causes a computer to execute the steps of: determining a forecast value of power demand for each of a number of defined regions in a region where power is exchanged, using actual values of past power demand; correcting the forecast value using the demand impact information and a time difference between the reference time and the time of power demand, which is set according to the correlation between the power demand of the region in the demand impact information, and the time difference between the reference time and the time of power demand, which is set according to the demand impact information and the power demand of the region; and managing the supply and demand of the entire region using the corrected forecast value.

Furthermore, AEMS1 may be realized by one computer system or by multiple computer systems. For example, AEMS1 may be realized by a cloud system.

BEMS 2-1 to 2-3 are also realized by computer systems. The communication unit 21 shown in FIG. 1 is realized by, for example, the communication unit 105 shown in FIG. 2. The demand forecasting unit 22 shown in FIG. 1 is realized by the control unit 101 executing a program. The memory unit 23 shown in FIG. 1 is realized by the memory unit 103 shown in FIG. 2. Furthermore, BEMS 2-1 to 2-3 may be realized by one computer system or multiple computer systems. For example, BEMS 2-1 to 2-3 may be realized by a cloud system.

Next, we will explain the method for correcting electricity demand in the embodiment. First, we will explain the difference in daily changes in electricity demand depending on the business type.

Figure 3 is a schematic diagram showing changes in electricity demand in an office. Figure 3 shows a schematic diagram of daily changes in electricity demand in the summer in an office with set working hours. As shown in Figure 3, electricity demand is high during working hours, and as air conditioning equipment is used as the temperature rises, electricity demand peaks around 1:00 p.m.

Figure 4 is a schematic diagram showing the change in electricity demand at a retail store. Figure 4 shows a schematic diagram of the daily change in electricity demand at a retail store in the summer. As shown in Figure 4, electricity demand is high during business hours. Also, compared to the electricity demand after working hours in the office shown in Figure 3, the electricity demand at the retail store after business hours drops sharply.

Figure 5 is a schematic diagram showing the change in electricity demand in a hotel. Figure 5 shows a schematic diagram of the daily change in electricity demand in a hotel in the summer. As shown in Figure 5, the change in electricity demand in a hotel is small throughout the day. This is because, while offices and retail stores have more guests during the day, hotels have more guests in the mornings and evenings than during the day, so the increase in electricity demand due to the use of air conditioning equipment and the like during the day is smaller than in offices and retail stores.

As such, daily changes in electricity demand differ depending on the type of business. Furthermore, the impact of social conditions due to infectious disease countermeasures such as teleworking, business voluntary restraints, and self-restraint on going out, as well as weather warnings, is also thought to differ depending on the type of business. For this reason, in this embodiment, a correction method is defined for each type of business.

Figure 6 is a diagram showing an example of the degree of impact of social conditions on electricity demand due to infection control measures for each business type. As shown in Figure 6, the implementation rate of telework for each business type and the demand fluctuation factors that are fluctuation factors for electricity demand are shown. This implementation rate of telework may be obtained by using publicly available statistical information, or, if there are multiple tenants of the same business type in a building in the area, information may be collected from these tenants.

The demand fluctuation factors indicate factors that affect the fluctuation of demand for each business type. In offices, the number of employees (the number of employees working in the building) is a factor of demand fluctuation, in hotels, the number of guests is a factor of demand fluctuation, and in retail stores, food supermarkets, medical institutions, and restaurants, the number of visitors is a factor of demand fluctuation. Influence level #1 shown in Figure 6 indicates the degree of influence on electricity demand due to telework, which is an example of the fluctuation of demand fluctuation factors, and influence level #2 indicates the degree of influence of refraining from going out. Regarding influence level #1, the number of employees in office buildings fluctuates, and since the number of employees is large, the influence of the fluctuation in the number of employees on electricity demand is large. In hotels, the number of guests staying overnight for business purposes increases or decreases due to telework, and the influence of the increase or decrease in the number of guests on electricity demand is large. In addition, in retail stores, food supermarkets, medical institutions, and restaurants, the number of visitors increases or decreases as telework increases, but for retail stores, food supermarkets, and medical institutions, although the electricity demand fluctuates due to the number of visitors, the proportion of electricity consumption by fixed equipment is higher than that of restaurants, and the influence of the number of visitors is small. In restaurants, the impact of teleworking is moderate because the increase or decrease in the number of customers has a large impact on electricity demand. In this way, the fluctuation factors and the degree of impact of the fluctuation factors on electricity demand differ depending on the business type, so in this embodiment, the fluctuation factors are estimated for each business type.

As shown in impact level #2 in Figure 6, the impact of voluntary stay-at-home orders will be large for offices, hotels, retail stores, supermarkets, medical institutions, and restaurants. On the other hand, there are business types that are easily affected by crowds and business types that are not, and when voluntary stay-at-home orders are planned based on wide-area information, AEMS1 makes corrections by business type. For example, each tenant is classified into business type #1, business type #2, and business type #3. Business type #1 is offices, hotels, etc. that will remain open even if voluntary stay-at-home orders are implemented, and business type #2 is retail stores, supermarkets, medical institutions, restaurants, etc. that will remain open even if voluntary stay-at-home orders are implemented. Business type #3 is a tenant that will be closed due to voluntary stay-at-home orders. Note that in order to distinguish between business types #1 to #3, it is necessary to know the business hours and whether or not the business will be closed, but this information can be obtained from business hours information as described below.

For example, business type #1 is a business type that continues to operate in the same manner as normal even if there is self-restraint, and AEMS1 sets a larger correction amount according to the foot traffic flow when making corrections in response to information indicating foot traffic flow, such as the number of passengers boarding and alighting at stations in the second information (wide area information), than for business type #2. Business type #2 is a business type that changes significantly from normal times, and AEMS1 sets a smaller correction amount according to the foot traffic flow when making corrections in response to information indicating foot traffic flow in the second information than for business type #1. Business type #3 is a business type that is closed, and since there are few fluctuations due to foot traffic if power corrections are made during closure taking into account business hours, i.e. using the first information, corrections that take into account the second information are not necessary.

In this embodiment, when the AEMS1 estimates that the value of the demand fluctuation factor exemplified above has changed from the date and time corresponding to the actual value used in the prediction by the demand forecasting unit 22 of each BEMS 2-1 to 2-3, the AEMS1 corrects the power demand based on the amount of change. In this embodiment, at least one of the first information and second information described above is used as information for grasping the change in the value of the demand fluctuation factor.

Here, an example of the first information used to calculate the correction information in this embodiment will be described. FIG. 7 is a diagram showing business hours information, which is an example of the first information. Although not shown in FIG. 7, the business hours information is associated with a date. In the example shown in FIG. 7, business hours are divided into standard business hours and actual business hours. Also, in FIG. 7, which of the above-mentioned business types #1 to #3 each tenant corresponds to is shown on the right side. The actual business hours store the latest information of the associated date. For example, if the business hours scheduled on the day before the demand occurs are changed on the day, the actual business hours in the business hours information are updated on the day. When the business hours information is updated, BEMS2-1 transmits the updated business hours information to AEMS1. BEMS2-1 acquires the business hours information shown in FIG. 7 from, for example, the building security management system 3. As shown in FIG. 7, the business hours information shows the business hours of each tenant in the building managed by BEMS2-1. Note that, if the tenant is an office, the business hours are set to, for example, regular working hours. The standard business hours can be used as an estimate of the business hours corresponding to the actual demand value used for estimation in the demand forecasting unit 22 of BEMS2-1.

As described above, the actual value of the power demand used in the demand forecasting unit 22 for the demand forecast is treated as the date and time when no social changes such as infectious disease countermeasures have occurred, so it is assumed that each tenant was affected by the standard business hours on these dates and times. Therefore, if the standard business hours and the actual business hours match, corrections due to changes in business hours do not need to be made to the power demand. When corrections are made using business hours, for example, the difference between the actual value of the power demand during closure and the actual value of the power demand during normal times for each tenant is calculated in advance, and this difference is stored in the storage unit 15 as correction information for each tenant and time period, so that the AEMS1 subtracts this difference from the predicted value of the power demand for each tenant when the tenant is closed. The unit of this time period may be 30 minutes, which is the calculation unit of the predicted value described above, or may be a multiple unit time period such as one hour. In addition, for example, when business hours change, a predetermined value is subtracted from the predicted value for the time period during which the business hours are shortened and no longer open, for each tenant and time period.

The above correction method is an example in which the correction information is directly determined, but the analysis unit 12 of the AEMS1 may use the first information, which is the standard business hours and actual business hours, and the actual power demand value, to calculate the actual business hours and the value to be increased or decreased from the predicted power demand value for each tenant by regression analysis, machine learning, etc. As described above, the predicted power demand value is calculated from the actual value of business during standard business hours. In addition, a correction coefficient used to calculate the predicted power demand value may be calculated by regression analysis, machine learning, etc., not for each tenant but for each business type.

Figure 8 is a diagram showing the attendance status, which is an example of the first information. The attendance status is information indicating the working style of an employee, and although not shown in Figure 8, it is associated with a date. The BEMS 2-1 acquires the attendance status shown in Figure 8 from, for example, the tenant's work management system 4. In the example shown in Figure 8, working styles such as normal attendance, telecommuting, and staggered attendance are shown for each employee. Normal attendance indicates that the employee comes to work in the area of each tenant, such as an office in a building. The attendance status is provided to the BEMS 2-1 from the work management system 4, for example, on the day before the demand occurs. The attendance status may be updated on the day. As shown in Figure 6, when the tenant is an office, the number of people present in the building depends on the attendance status, but when the tenant is a food supermarket, restaurant, etc., the number of visitors has a greater impact than the attendance status of employees. For this reason, the BEMS 2-1 may acquire the attendance status from tenants in business types that have a large impact on the power demand due to the attendance status, such as offices, and may not acquire the attendance status from tenants in business types that have a small impact on the power demand due to the attendance status.

Based on the attendance situation, the number of employees, which is a factor of fluctuation in office demand shown in Figure 6, can be calculated. Therefore, the predicted value of electricity demand can be corrected using the ratio of the number of employees in normal times, i.e., before changes in social conditions such as infection control measures, to the calculated number of employees. For example, correction can be made by multiplying the predicted value of electricity demand by (calculated number of employees / number of employees in normal times). In this case, the formula or table for this calculation is stored in the memory unit 15 of the AEMS1 as correction information.

In addition, the analysis unit 12 of the AEMS 1 may calculate the coefficients a and b in the following formula (1) by regression analysis and machine learning using the number of employees after the introduction of teleworking and the corresponding actual value of power demand, and store this coefficient as correction information in the storage unit 15. That is, a and b are learned so that the difference between the corrected predicted value calculated by the following calculation formula and the actual value of power demand is minimized. Note that the calculation formula is not limited to the form of the following formula (1), and may be any formula that uses the calculated number of employees to obtain a corrected predicted value. The same values of a and b may be used for 24 hours, or a and b may be determined for each time zone in 30-minute units. Also, a and b may be determined for each season, day of the week, etc. Note that this correction is performed, for example, taking into account the time difference described below. For example, the correction based on the attendance status considers an event that occurs at the start time of regular work hours, and the correction is performed taking into account the time difference from the start time. For example, if the time difference is one hour and the start time is 9:00, the following correction is performed on the predicted value for 10:00.
Corrected forecast value = Uncorrected forecast value × a × Number of employees + b ... (1)

The first information may be an image of a surveillance camera or information obtained from an image of a surveillance camera. BEMS2-1 acquires the image of the surveillance camera from, for example, the building security management system 3. FIG. 9 is a diagram showing the relationship between the image of the surveillance camera and the prediction of the number of people in a certain floor of a building. As shown in FIG. 9, a surveillance camera is installed at an entrance on a certain floor of a building managed by BEMS2-1. BEMS2-1 or AEMS1 calculates the number of people passing through the entrance per hour by analyzing the image acquired by the surveillance camera. Note that a general method for detecting people in an image can be used by image processing, so a detailed description is omitted. When BEMS2-1 transmits the image of the surveillance camera to AEMS1 as the first information, the analysis unit 12 of AEMS1 calculates the number of people passing through the entrance per hour using the image. In addition, BEMS2-1 may be provided with an image processing unit (not shown), and the image processing unit may calculate the number of people passing through the entrance per hour using the image. Alternatively, if the building security management system 3 has a function for calculating the number of people passing through an entrance/exit per hour, the BEMS 2-1 may acquire the number of people passing through an entrance/exit per hour calculated by the building security management system 3 instead of an image as the first information.

The number of people passing through the entrance and exit per hour (hereinafter also simply referred to as the number of people passing through) calculated from the surveillance camera image shown in FIG. 9 does not represent the actual number of people (active users) staying on the floor of the building, but it is considered that the number of people passing through increases as the number of people staying on the floor of the building increases, and the number of people passing through decreases as the number of people staying on the floor of the building decreases. Therefore, the power demand on the floor of the building in normal times and the ratio between the number of people passing through in normal times and the number of people passing through on the day of the power demand prediction can be used to correct the power demand. For example, if the number of active users is 50 and the number of people passing through is 100 in normal times, and if the number of people passing through on the day of the prediction is 50, the estimated number of active users will be 25, which is half of the normal number. Therefore, for example, the power demand is also estimated to be halved.

Figure 10 is a diagram showing an example of the correction of electricity demand based on the number of passersby recognized from the surveillance camera image shown in Figure 9. Figure 10 shows the average number of passersby recognized from the surveillance camera image shown in Figure 9, the number of active users, and electricity demand, respectively, during normal times, rainy weather, and telework. It is assumed that some of Tenants A to D shown in Figure 9 are offices, and the others are retail stores. For this reason, as shown in Figure 10, the number of passersby decreases during telework, i.e., when office tenants are teleworking, and the number of passersby decreases during rainy weather (when telework is not being performed), due to the decrease in the number of customers at retail stores.

Figure 10 shows an example of correction when a formula or table for correction is directly set as correction information, and assumes that power demand is directly proportional to the number of passersby. As shown in Figure 10, in normal times, when the number of active users is 50, the power demand is 10 kWh, and the number of passersby is 100, if the forecast target day is rainy and the average number of passersby recognized from the surveillance camera image is 50, the predicted value of power demand is corrected to 50 x (50/100) = 50 x (1/2) kWh. Similarly, in teleworking, if the average number of passersby recognized from the surveillance camera image is 10, the predicted value of power demand is corrected to 50 x (10/100) = 50 x (1/10) kWh.

In FIG. 10, rainy weather and teleworking are described separately to correspond to the phenomenon, but the power demand may be corrected based on the ratio between the number of passersby in normal times and the number of passersby on the day for which the power demand is predicted, without distinguishing between the causes. That is, the calculation formula of the following formula (2) or a table corresponding to the following formula (2) is stored in the storage unit 15 of the AEMS 1 as correction information. The power demand in normal times is a predicted value of the power demand transmitted from the BEMS 2-1 to 2-3. The reference number of passersby is a predetermined constant, and is set as the number of passersby in normal times, for example. This correction is also performed, for example, taking into account the time difference described below.
Corrected forecast value = Uncorrected forecast value × (number of pedestrians on the forecast target day / reference number of pedestrians) ... (2)

In addition, it is also possible to obtain actual values of the number of pedestrians and electricity demand under normal conditions for each weather condition, such as rainy or sunny, and determine the standard number of pedestrians in the above formula (2) for each weather condition.

In the above example, the correction information is directly set, but the analysis unit 12 of the AEMS 1 may use the surveillance camera image and the actual value of the power demand, which are the first information, to calculate the number of passersby recognized from the surveillance camera and the correction coefficient to be multiplied by the predicted value of the power demand by regression analysis, machine learning, or the like. For example, the analysis unit 12 may calculate the number of passersby per hour by performing image processing on the surveillance camera image acquired from the BEMS 2-1, and the number of passersby and the predicted value and actual value of the power demand at the date and time corresponding to the image may be set as one data set, and the coefficients α and β in the following formula (3) may be determined by regression analysis, machine learning, or the like using a large number of data sets acquired from the BEMS 2-1. In this case, the coefficients α and β are stored in the storage unit 15 as correction information.
Corrected predicted value = Predicted value before correction × α × Number of pedestrians + β ... (3)

The values of α and β may be the same for 24 hours, or may be determined for each time period in 30-minute units.Also, α and β may be determined for each season, day of the week, etc.
Furthermore, as described above, when the BEMS 2-1 or the building security management system 3 calculates the number of passersby from the images from the surveillance camera, the BEMS 2-1 transmits the number of passersby as the first information to the AEMS 1, and the analysis unit 12 can similarly calculate the correction information. Note that the calculation formula is not limited to the form of the above formula (3), and any formula that can obtain a corrected predicted value using the number of passersby recognized by the surveillance camera may be used.

Although the above are examples of the first information, the first information is not limited to these, and can be set appropriately depending on the business type of the tenants residing in the buildings managed by each BEMS 2-1 to 2-3, information that can be obtained from the tenants, and information that can be obtained in the building. In addition, even if the first information uses entrance and exit management information for each room in the building, the number of times the elevator goes up and down, and the number of people entering the building (the number of people passing through the entrance gate), the power demand can be corrected in a similar manner depending on the difference in the number of people compared to normal times. In addition, since the number of times the building doors and windows are opened and closed affects the efficiency of the air conditioning equipment, the prediction accuracy can be improved by making corrections such as increasing the predicted value when the number of times the building doors and windows are opened and closed is high.

Next, the second information will be described. FIG. 11 is a diagram showing an example of the relationship between the number of passengers boarding and alighting at a station, which is an example of the second information, and the power demand. The second information is information obtained from the external system 5 as described above. The bar graph shown in FIG. 11 shows the number of passengers boarding and alighting at the nearest station to the building managed by the BEMS 2-1 in each time period. The power demand 201 shows the power demand in the summer of the building managed by the BEMS 2-1. Here, it is assumed that most of the tenants in the building managed by the BEMS 2-1 are offices, so that a peak value 203 of the number of passengers boarding and alighting occurs before 9:00, which is the start time of the office work. If the number of passengers boarding and alighting at this time is large, the power demand 201 will also increase. Therefore, as with the first information described above, the influence of the number of passengers boarding and alighting at the station can be reflected in the correction by determining correction information in advance for the number of passengers boarding and alighting at the station, or by determining correction information by regression analysis, machine learning, or the like.

Meanwhile, while a peak value 203 in the number of passengers getting on and off occurs before 9:00, the peak value 202 of the office's electricity demand occurs around 15:00. The peak value 202 of the office's electricity demand depends on the number of employees who come to the office. Therefore, the number of passengers getting on and off between 7:00 and 8:00 will affect the peak value 202 of the office's electricity demand around 15:00.

It is also possible that the peak value 204 of the rate of increase in the number of passengers boarding and alighting at the station in the morning hours shown in FIG. 11 is correlated with the electricity demand. For this reason, like the peak value 203 of the number of passengers, the peak value 204 of the rate of increase before working hours may be used to correct the electricity demand. The peak value 203 of the number of passengers boarding and alighting and the peak value 204 of the rate of increase in the number of passengers are, in a broad sense, information indicating the number of passengers boarding and alighting at the station. In this way, the second information used to correct the electricity demand may be the number of passengers boarding and alighting at the station in each time period, the peak value 203 of the number of passengers boarding and alighting, or the peak value 204 of the rate of increase in the number of passengers.

As such, depending on the information, there may be a time difference between the time corresponding to the value of that information and the time when that value has a noticeable effect on electricity demand. The number of passengers boarding and alighting at the above-mentioned stations has a relatively large time difference, but the above-mentioned business information, etc., has an effect of the value with almost no time difference. For this reason, in this embodiment, the analysis unit 12 calculates the correlation for each time difference between the information and electricity demand for each piece of first information and second information.

Figure 12 is a schematic diagram showing correlation values for each time difference of each piece of information. The horizontal axis of Figure 12 indicates the time difference. Correlation values 301 to 304 indicate correlation values between information on different types of information and power demand. For example, the analysis unit 12 collects multiple pieces of building power demand and station passenger numbers for the same season and the same day of the week, sets the peak number of passengers at the station to x, calculates the time when the number of passengers at the station peaks, and sets the building power demand at a time ΔT after the calculated time to y. The analysis unit 12 then uses multiple sets of x and y to find a correlation value between x and y. This is repeated to find correlation values by sequentially changing the time difference ΔT from, for example, 0 to 10 hours. Correlation values are similarly found for other pieces of information to obtain correlation values 301 to 304 as shown in Figure 12. For each piece of information, the information is used to correct the time period to which the time difference found in this way is given. For example, in the case where the correlation between the number of passengers boarding and alighting at a station in the morning and the value of the power demand five hours later is the highest, the morning peak value of the day is reflected in the power demand five hours later. Specifically, for example, machine learning is performed to obtain the above correction information by reflecting the time difference obtained in this manner. For example, when performing machine learning, a coefficient for correction is obtained by using the number of passengers boarding and alighting at time t, the actual value of the power demand five hours later, and the predicted value of the power demand five hours later as a data set. Then, when performing the correction, this correction is reflected after the power demand five hours later. That is, the correction unit 13 of this embodiment corrects the predicted value of the power demand using the time difference at which the impact of the demand impact information on the power demand of the building becomes large, and the demand impact information, among the time differences between the reference time set in response to the demand impact information and the time of the power demand. Note that in the above example, the analysis unit 12 obtains the correlation for each time difference (the correlation between the demand impact information and the power demand) and calculates the time difference at which the correlation is high, but the time difference at which the correlation is high may be input by an operator, etc.

When the power demand five hours from now is corrected using the peak number of passengers, the power demand for a specified period before and after five hours may be increased or decreased at the same rate as after five hours. For example, the power demand for a period of one hour before and after five hours may be increased or decreased at the same rate as after five hours. In this way, when making corrections taking into account a time difference, corrections may be made not only for that point in time but also for the periods before and after. These periods before and after are also determined in advance by analysis.

By using information with a high correlation and a large time difference, the power demand for that day can be reflected in the forecast value with plenty of time. In particular, when it is desired to grasp the peak value of the power demand for a day, it is possible to reflect it in the forecast well before the time period when the peak value occurs, which makes it easier to reflect it in the power procurement plan and the power suppression plan. As described above, the time of the information when calculating the time difference for information such as the attendance status may be, for example, the start time of the workday, or a time a certain number of hours after the start time of the workday, or a time a certain number of hours after the start time of the workday. In addition, for information based on a measurement value such as the number of passersby obtained from a surveillance camera image, the reference time when considering the time difference is, for example, the time when the measurement value was measured. For example, in the case of the number of passersby obtained from a surveillance camera image, it is the time when the image was obtained by the surveillance camera. For values within a period such as the number of passersby per hour, the reference time may be the start time, the center time, or the end time of the period, as long as it is a time within the period.

Figure 13 is a diagram showing the influence of weather information, which is another example of the second information, on the power demand. The bar graph in Figure 13 shows the number of passengers at stations on days when weather warnings such as heavy rain warnings and storm warnings were issued. In the example of the upper diagram in Figure 13, warnings are issued for a long time period from about 6:00 to about 20:00, and in the example of the lower diagram in Figure 13, warnings are issued only in the morning. The power demand shows the normal power demand 201 shown in Figure 11, power demand 205 shows the power demand when a warning is issued for a long time period from about 6:00 to about 20:00, and power demand 208 shows the power demand when a warning is issued only in the morning. As shown in the upper diagram in Figure 13, when a warning is issued for a long time period from about 6:00 to about 20:00, the peak value 207 of the increase rate of the number of passengers is reduced compared to the peak value 204 of the increase rate in Figure 11, and the peak value 206 of the power demand is also reduced compared to the peak value 202 of the power demand 201 in Figure 11. On the other hand, as shown in the lower diagram of FIG. 13, if a weather warning is issued only in the morning, the start time of work will be shifted, but the peak value 209 itself, as shown in power demand 208, may not change much from normal times. In this way, if a weather warning is issued only in the morning, the peak of the increase rate of the number of passengers boarding and alighting at the station will appear later than in the example shown in FIG. 11. For this reason, if a weather warning is issued in the morning and then lifted near the start of work, the time difference between the number of passengers boarding and alighting and the peak value of power demand may differ from normal times.

As such, since the power demand may change depending on the time period during which a weather warning is issued, the analysis unit 12 may learn the time series data of the power demand by machine learning in association with the time period during which the warning is issued (e.g., 6:00 to 8:00, 6:00 to 20:00, etc.). For example, the analysis unit 12 performs machine learning using a data set of the time periods during which past weather warnings were issued, time series data of the predicted power demand, and time series data of the daily change in the actual power demand, to generate a learned model as correction information for correcting the predicted power demand for each time period during which a weather warning is issued.

The second information may be mobile phone location information or business information from corporate public relations etc. When using mobile phone location information, the AEMS1 estimates the number of people in the building using the mobile phone location information, in the same way as the number of passersby recognized by the surveillance camera, and corrects the power demand using the estimated number of people.

Although examples of the first information and the second information have been described above, the predicted value of the power demand may be corrected by one of the first information and the second information, or the predicted value of the power demand may be corrected by two or more pieces of information. For example, the analysis unit 12 may generate a trained model for correcting the predicted value of the power demand by machine learning for each tenant or each business type, using the first information, the second information, the actual value of the power demand, and the predicted value of the power demand, taking into account the time difference.

Next, a correction information generation process procedure when the analysis unit 12 generates correction information will be described. FIG. 14 is a flowchart showing an example of the correction information generation process procedure of this embodiment. The AEMS 1 also stores the pre-correction predicted values acquired from the BEMS 2-1 to 2-3 in the memory unit 15 as past predicted values. The analysis unit 12 of the AEMS 1 acquires first information from the BEMS 2-1 to 2-3 via the communication unit 14 (step S1) and acquires second information from the external system 5 (step S2). The analysis unit 12 also acquires performance information from the BEMS 2-1 to 2-3 via the communication unit 14 (step S3). The first information and second information acquired in steps S1 and S2 include information on the date and time corresponding to the performance information. The performance information is the actual value of the power demand measured in each building managed by the BEMS 2-1 to 2-3, and preferably includes the date and time when the change in the social situation due to the infectious disease is reflected and the date and time when a weather warning is issued.

The analysis unit 12 generates corrected information (corrected model) using the first information, the second information, the predicted value before correction, and the actual value of the power demand included in the actual information for the same or similar date and time as a data set (step S4). The corrected information is, for example, a trained model generated by machine learning as described above.

Next, the correction of the power demand in the AEMS 1 of this embodiment will be described. FIG. 15 is a flowchart showing an example of a correction process procedure of the power demand in the AEMS 1 of this embodiment. The AEMS 1 acquires the demand amount, i.e., the predicted value of the power demand from the BEMS 2-1 to 2-3 (step S11), acquires the first information from the BEMS 2-1 to 2-3 (step S12), and acquires the second information from the external system 5 (step S13). In detail, the communication unit 14 receives the predicted value of the power demand and the first information from the BEMS 2-1 to 2-3 and stores them in the memory unit 15, and acquires the second information from the external system 5 and stores them in the memory unit 15. Next, the AEMS 1 corrects the predicted value using the first information, the second information, and the correction information (correction model) (step S14). In detail, the correction unit 13 corrects the predicted value stored in the memory unit 15 using the first information, the second information, and the correction information stored in the memory unit 15. As described above, both the first information and the second information may be used, or only one of them may be used.

Note that the accuracy of the correction can be improved by making the correction while taking into account various pieces of information as the first information and the second information, but on the other hand, as the amount of information used increases, the capacity for data storage and the processing time increase. For this reason, AEMS1 may select and use the information that has the greater impact on power demand from the first information and the second information. Figure 16 is a schematic diagram showing an example of the locations of buildings and stations. As shown in Figure 16, it is assumed that buildings #1 to #3 exist within an area managed by AEMS1, building #1 is close to station S, building #2 is close to station T, and building #3 is about the same distance from station S and station T. If buildings #1 to #3 are managed by BEMS 2-1 to 2-3, respectively, it is considered that the number of passengers boarding and alighting at station S is desirably used to correct the predicted value of the power demand of BEMS 2-1, the number of passengers boarding and alighting at station T is desirably used to correct the predicted value of the power demand of BEMS 2-2, and both the number of passengers boarding and alighting at station T and the number of passengers boarding and alighting at station S are desirably used to correct the predicted value of the power demand of BEMS 2-3. For example, the analysis unit 12 may obtain a correlation between the number of passengers boarding and alighting at each station and the power demand, and remove information with a weak correlation from the input. Alternatively, correction may be actually performed using both a model that takes each piece of information into account and a model that does not take each piece of information into account, the degree of influence of the information may be evaluated, and information with a small degree of influence may be removed.

In the above example, AEMS1 is provided with the correction unit 13, but each BEMS2-1 to 2-3 may be provided with the analysis unit 12 and the correction unit 13, and each BEMS2-1 to 2-3 may correct the predicted value. Also, two-stage correction may be performed, in which each BEMS2-1 to 2-3 performs correction using the first information and AEMS1 performs correction using the second information. Also, in the above example, AEMS1 is provided with the analysis unit 12, but a device other than AEMS1 may be provided with the analysis unit 12, and AEMS1 may obtain correction information from the other device. Also, a device other than AEMS1 may be provided with the analysis unit 12, and each BEMS2-1 to 2-3 may be provided with the correction unit 13 and obtain correction information from the other device.

As described above, in this embodiment, the predicted value of the power demand is corrected using the demand impact information, which is information that may affect the power demand, and the time difference at which the correlation between the demand impact information and the power demand is high. This makes it possible to improve the accuracy of the power demand forecast.

The configurations shown in the above embodiments are merely examples, and may be combined with other known technologies, or the embodiments may be combined with each other. In addition, parts of the configurations may be omitted or modified without departing from the spirit of the invention.

1 AEMS, 2-1 to 2-3 BEMS, 3 Building security management system, 4 Work management system, 5 External system, 11 Supply and demand management unit, 12 Analysis unit, 13 Correction unit, 14, 21 Communication unit, 15, 23 Memory unit, 16 Display unit, 22 Demand forecast unit.

Claims (16)

A regional energy management device that manages the supply and demand of electricity in the entire region where the power exchange is carried out;
A plurality of facility energy management devices each managing power in a region defined in the area;
Equipped with
The facility energy management devices calculate a predicted value of the power demand of the corresponding area using a past actual value of the power demand, and transmit the predicted value to the local energy management device;
the regional energy management device corrects the predicted value using the demand impact information and a time difference between a reference time set in response to demand impact information, which is information that may affect the power demand of the region, and the time of the power demand, the time difference being set based on a correlation between the demand impact information and the power demand of the region, and uses the demand impact information to manage supply and demand throughout the region;
The demand impact information includes the number of passengers boarding and alighting at a station,
The regional energy management device uses the acquired demand impact information and power demand to calculate a correlation value between the demand impact information and the power demand for each time difference between the reference time and the time of the power demand, and uses the calculated correlation value to set the time difference to be used for the correction of the predicted value. A regional energy management device that manages the supply and demand of electricity in the entire region where the power exchange is carried out;
A plurality of facility energy management devices each managing power in a region defined in the area;
Equipped with
The facility energy management devices calculate a predicted value of the power demand of the corresponding area using a past actual value of the power demand, and transmit the predicted value to the local energy management device;
the regional energy management device corrects the predicted value using the demand impact information and a time difference between a reference time set in response to demand impact information, which is information that may affect the power demand of the region, and the time of the power demand, the time difference being set based on a correlation between the demand impact information and the power demand of the region, and uses the demand impact information to manage supply and demand throughout the region;
The demand impact information includes a weather warning;
The regional energy management device uses the acquired demand impact information and power demand to calculate a correlation value between the demand impact information and the power demand for each time difference between the reference time and the time of the power demand, and uses the calculated correlation value to set the time difference to be used for the correction of the predicted value. The local energy management system according to claim 1 or 2, characterized in that the area is a building. The regional energy management system according to claim 3, characterized in that the demand impact information includes first information acquired by the facility energy management device in the building and second information acquired from an external system outside the regional energy management system. The local energy management system according to claim 4, characterized in that the first information includes information indicating the working patterns of employees of tenants residing in the building. The local energy management system according to claim 4 or 5, wherein the first information includes information indicating business hours of tenants in the building. The local energy management system according to any one of claims 4 to 6, characterized in that the first information includes entry and exit management information for each room in the building. The local energy management system according to any one of claims 4 to 7, characterized in that the first information includes the number of times the doors of the building are opened and closed. The local energy management system according to any one of claims 4 to 8, characterized in that the first information includes the number of times the windows of the building are opened and closed. The local energy management system according to any one of claims 4 to 9, characterized in that the first information includes the number of times the elevator in the building goes up and down. The local energy management system according to any one of claims 4 to 10, characterized in that the second information includes mobile phone location information. The local energy management system according to any one of claims 1 to 11, characterized in that the local energy management device uses a correction model learned by machine learning using the demand impact information and the actual values to infer a correction value to be used in the correction of the predicted value, and corrects the predicted value using the correction value. A regional energy management device that manages the supply and demand of electricity in an entire region where power interchange is performed,
a communication unit that receives, from a plurality of facility energy management devices that manage the power of corresponding areas defined in the region, predicted values of power demand for the corresponding areas calculated using actual values of power demand in the past, and receives demand impact information that may affect the power demand for the corresponding areas;
a correction unit that corrects the predicted value by using a time difference between a reference time set corresponding to the demand impact information and a time of the power demand, the time difference being set based on a correlation between the demand impact information and the power demand in the region, and the demand impact information;
a supply and demand management unit that manages supply and demand in the entire area by using the corrected predicted value;
Equipped with
calculating a correlation value between the demand impact information and the power demand for each time difference between the reference time and the time of the power demand, using the demand impact information and the power demand that have been acquired, and setting the time difference to be used for the correction of the predicted value, using the calculated correlation value;
A regional energy management device, wherein the demand impact information includes the number of passengers boarding and alighting at stations. A facility energy management device that manages the power of a respective defined area in a region where power interchange is performed,
a demand forecasting unit that obtains a forecast value of the power demand of the corresponding area by using a past actual value of the power demand;
a correction unit that corrects the predicted value using a time difference between a reference time and a time of power demand that is set based on a correlation between the demand impact information and the power demand of the region, the time difference being set in accordance with demand impact information that is information that may affect the power demand of the region, and the demand impact information;
Equipped with
calculating a correlation value between the demand impact information and the power demand for each time difference between the reference time and the time of the power demand, using the demand impact information and the power demand that have been acquired, and setting the time difference to be used for the correction of the predicted value, using the calculated correlation value;
A facility energy management device, wherein the demand impact information includes the number of passengers boarding and alighting at stations. A step of calculating a predicted value of power demand for each of a plurality of defined regions in a region where power interchange is performed, using a past actual value of power demand;
a step of correcting the predicted value using the demand impact information and a time difference between a reference time set in accordance with demand impact information, which is information that may affect the power demand of the region, and the time of the power demand, the time difference being set based on a correlation between the demand impact information and the power demand of the region, and the demand impact information, and managing the supply and demand of the entire region using the corrected predicted value;
Including,
The demand impact information includes the number of passengers boarding and alighting at a station,
A demand forecasting method characterized in that the time difference used for the correction of the predicted value is calculated using the demand impact information and the electricity demand that have already been acquired, and a correlation value between the demand impact information and the electricity demand for each time difference between the reference time and the time of the electricity demand. A step of calculating a predicted value of power demand for each of a plurality of defined regions in a region where power interchange is performed, using actual values of past power demand;
a step of correcting the predicted value using the demand impact information and a time difference between a reference time set in accordance with demand impact information, which is information that may affect the power demand of the region, and the time of the power demand, the time difference being set based on a correlation between the demand impact information and the power demand of the region, and the demand impact information, and managing the supply and demand of the entire region using the corrected predicted value;
Run the following on your computer:
The demand impact information includes the number of passengers boarding and alighting at a station,
A demand forecasting program characterized in that the time difference used for the correction of the predicted value is calculated using the demand impact information and the electricity demand that have already been acquired, and a correlation value between the demand impact information and the electricity demand for each time difference between the reference time and the time of the electricity demand.

Images