JP7343857B2 - Analysis equipment - Google Patents

Description

Application of Article 30, Paragraph 2 of the Patent Act July 11, 2018 Published in the 2018 Society of Air Conditioning and Sanitation Engineers Conference Proceedings Volume 9 August 29, 2018 2018 Air Conditioning and Sanitation Engineering Society Conference Proceedings Volume 9 Published in the Society of Engineering Conference Conference Proceedings (DVD-R) September 13, 2018 Published at the 2018 Air Conditioning and Sanitation Engineering Society Conference (Nagoya) December 3, 2018 Asim2018 IBPSA Asia Conference Report Published in the proceedings December 3, 2018 Published in Asim2018 IBPSA Asia Conference January 29, 2019 Published in the 35th Energy Systems, Economy, and Environment Conference Lecture Proceedings Pages 592-597 2019 January 30, 2019 Published at the 35th Energy Systems, Economy, and Environment Conference March 5, 2019 Proceedings of the Society of Air Conditioning and Sanitary Engineers Vol. 44 No. Published on 264

The present disclosure relates to a technique for analyzing power consumption status of a building.

Conventionally, techniques have been developed to analyze the power consumption of each building based on the power consumption of the building. For example, Japanese Unexamined Patent Publication No. 2008-298375 (Patent Document 1) discloses an air conditioning load calculation system that calculates the air conditioning load of a building based on the history of energy usage in the building. The air conditioning load calculation system selects a second predetermined period in which the amount of energy used is the least in the first predetermined period, based on data related to the amount of energy used at each time in the first predetermined period. The second predetermined period is, for example, one week. The air conditioning load calculation system calculates the energy base usage amount for the first predetermined period based on the energy usage amount for the second predetermined period. Energy base usage is calculated as energy usage excluding air conditioning load. The air conditioning load calculation system calculates the amount of energy used for air conditioning based on the energy base usage amount.

Japanese Patent Application Publication No. 2008-298375

In the technique described in Patent Document 1, the energy base usage amount is calculated based on the energy usage amount for the second predetermined period. Since the second predetermined period is short, for example, one week, the accuracy of the energy base usage is low. As a result, the accuracy of calculating the amount of energy used for air conditioning also decreases.

The present invention has been made in view of the above-mentioned problems, and its purpose is to provide a technique that allows accurate analysis of power consumption by application.

According to a certain aspect, an analysis device that analyzes the power consumption state of a building includes an acquisition section, a determination section, and an estimation section. The acquisition unit acquires power amount data indicating power consumption measured for each time zone of the building during the analysis target period. The determination unit determines whether each day included in the analysis target period is an operating day or a non-operating day of the air conditioner. The estimating unit estimates the power consumption of the building by use in the target time period of the target day included in the analysis target period based on non-working day data that is data on non-working days among the power amount data. The determination unit extracts, from the power amount data, data on days in which the maximum power consumption among the power consumption measured for each time period in one day is a lower predetermined number of days, as training data. For each day included in the analysis target period, the determination unit determines whether the day is a working day or a non-working day, based on the comparison result between the training data and the judgment target data that is the data for that day among the electric energy data. Determine whether

In the above analysis device, the determination unit creates test data, which is time-series data that combines the judgment target data after the training data, and calculates the change point score at each point in time in the test data using a change point detection algorithm. You may. The determination unit determines whether or not a first change point score at a time point corresponding to the training data among the test data is equivalent to a second change point score at a time point corresponding to the determination target data among the test data. , when the first change point score and the second change point score are equivalent, the day corresponding to the determination target data is determined to be a non-working day, and the first change point score and the second change point score are If they are not equivalent, it may be determined that the day corresponding to the determination target data is a working day. The test is, for example, a t-test.

In the above analysis device, the estimating unit may estimate the air conditioning power consumption for air conditioning in the target time zone on the target day to be 0 when the target day is determined to be a non-working day. Furthermore, when the target day is determined to be a working day, the estimation unit calculates the target power consumption during the target time period on the non-working day calculated from the non-working day data from the target power consumption measured during the target time period on the target day. The amount obtained by subtracting the representative value of the power consumption may be estimated as the air conditioning power consumption in the target time period on the target day.

In the above analysis device, the acquisition unit may further acquire outside temperature data indicating the outside temperature for each time zone in the area including the building during the analysis target period. The estimation unit may extract the minimum power consumption for each outside temperature range using the power amount data and the outside temperature data. The estimation unit further calculates a regression equation with the outside temperature as an explanatory variable and the extracted minimum power consumption as an objective variable, and based on the regression equation and the outside temperature in the target time period on the target day, The base power consumption during the target time period may be estimated.

In the above analysis device, the estimator calculates the representative value of the base power consumption in the target time period on the non-working day from the representative value of the power consumption in the target time period on the non-working day calculated from the non-working day data. The amount obtained by subtracting the amount may be estimated as the activity-based power consumption amount for uses caused by human activities during the target time period.

The above analysis device is designed to match the target power consumption and the total sum when there is a difference between the target power consumption and the sum of air conditioning power consumption, base power consumption, and activity power consumption. The apparatus may further include a correction unit that corrects at least one of the air conditioning power consumption and the activity power consumption.

In the above analysis device, an activity time period may be determined in advance, which is a time period during which human activities are scheduled to take place in the building. Then, when the target time period is included in the active time period and the target power consumption amount is larger than the total sum, the correction unit may increase the air conditioning power consumption amount by the difference. The correction unit calculates the air conditioning power consumption by the difference when the target time period is included in the activity time period, the target power consumption is smaller than the total, and the difference is less than or equal to the air conditioning power consumption. You can reduce it. The correction unit calculates the air conditioning power consumption when the target time period is included in the activity time period, the target power consumption is smaller than the total, and the difference exceeds the air conditioning power consumption by the excess amount. may be reduced to 0, and the activity power consumption may be reduced by the excess amount. The correction unit may increase the activity amount power consumption by the difference when the target time period is not included in the activity time period and the target power consumption amount is smaller than the total sum. If the target time period is not included in the activity time period, the target power consumption is smaller than the total, and the difference is less than or equal to the air conditioning power consumption, the correction unit calculates the air conditioning power consumption by the difference. You can reduce it. The correction unit calculates the air conditioning power consumption when the target time period is not included in the activity time period, the target power consumption is smaller than the total, and the difference exceeds the air conditioning power consumption by the excess amount. may be reduced to 0, and the activity power consumption may be reduced by the excess amount.

According to another aspect, an estimation method for estimating a power consumption state of a building includes first to third steps. The first step is to acquire power amount data indicating the power consumption measured for each time zone of the building during the analysis target period. The second step is a step of determining whether each day included in the analysis target period is an air conditioner working day or a non-working day. The third step is to estimate the power consumption by usage of the building during the target time period on the target day included in the analysis target period, based on the non-operating day data, which is data on non-working days among the power consumption data. It is a step. The determining step includes a step of extracting, as training data, data on days in which the maximum power consumption among the power consumption measured for each time period of the day is lower than the predetermined number of days from the power amount data, and analysis. For each day included in the target period, determine whether the day is a working day or a non-working day based on the comparison result between the training data and the judgment target data, which is the data for that day among the electric energy data. and a step of doing so.

In other aspects, the program causes a computer to perform each step of the analysis method described above.

According to the technology of the present disclosure, power consumption by application can be analyzed with high accuracy.

1 is a diagram schematically showing a system including an analysis device according to the present embodiment. FIG. 2 is a diagram showing the hardware configuration of an analysis device. FIG. 3 is a diagram showing the functional configuration of an analyzer. It is a flowchart which shows an example of the flow of analysis processing in an analyzer. It is a figure explaining the estimation method of base power consumption. It is a figure which shows the verification result of the estimation accuracy of the base power consumption estimated by the estimation part. It is a figure which shows the example of application of ChangeFinder to test data. It is a figure which shows the verification result of the estimation accuracy of the activity part power consumption estimated by the estimation part. FIG. 2 is a diagram illustrating a method for estimating power consumption for air conditioning. It is a figure which shows the verification result of the estimation accuracy of the air conditioning power consumption estimated by the estimation part. FIG. 7 is a diagram showing the verification results of the accuracy of decomposition of power consumption into three uses. FIG. 7 is a diagram showing average values of errors between estimated values and actual measured values of base power consumption, activity power consumption, and air conditioning power consumption.

Embodiments according to the present invention will be described below with reference to the drawings. In the following description, the same parts and components are given the same reference numerals, and overlapping descriptions will not be repeated. Note that the embodiments and modifications described below may be selectively combined as appropriate.

<A. System configuration>
With reference to FIG. 1, the configuration of a system including an analyzer according to this embodiment will be described. FIG. 1 is a diagram schematically showing a system including an analysis device according to this embodiment. As shown in FIG. 1, the system SYS includes an analysis device 100, a building 200, and a server device 300.

The building 200 is, for example, a small or medium-sized business building such as a store or an office. A smart meter 210 is installed in the building 200. The smart meter 210 measures and records the total power consumption (hereinafter simply referred to as "power usage") for each time period used by various devices 220 in the building 200. Specifically, power consumption is measured for each time period from 0:00 to 24:00 divided by a predetermined time length (for example, 30 minutes or 15 minutes).

The smart meter 210 communicates with the server device 300 and transmits power amount data indicating the measured power consumption to the server device 300.

The server device 300 accumulates power amount data received from the smart meter 210 of the building 200.

The analysis device 100 is connected to a network, and acquires power amount data of the building 200 from the server device 300 via the network. The analysis device 100 analyzes the power consumption state of the building 200 based on the power data. The analysis device 100 may generate proposal information regarding energy conservation of the building 200 using the analysis results. The generated proposal information is distributed to the administrator who manages the building 200 as appropriate. By checking the proposal information, the administrator can take appropriate actions to save energy in the building 200.

<B. Analyzer hardware configuration>
FIG. 2 is a diagram showing the hardware configuration of the analysis device. As shown in FIG. 2, the analysis device 100 includes, as main components, a processor 101 that executes a program, a ROM (Read Only Memory) 102 that stores data in a nonvolatile manner, and a A RAM (Random Access Memory) 103 that volatilely stores data entered or input via an input device, a hard disk (HDD) 104 that stores data nonvolatilely, and a communication IF (Interface) 105. , an operation key 106, a power supply circuit 107, and a display 108. Each component is connected to each other by a data bus. Note that the communication IF 105 is an interface for communicating with other devices.

Processing in the analysis device 100 is realized by each piece of hardware and a program 110 executed by the processor 101. Such a program 110 is stored in the HDD 104 in advance. However, the program 110 may be stored in another storage medium and distributed as a program product. Alternatively, the program 110 may be provided as a downloadable program product by an information provider connected to the Internet. Such a program 110 is stored in the HDD 104 after being read from the storage medium by a reading device or downloaded via the communication IF 105 or the like. The processor 101 reads the program 110 from the HDD 104 and executes the program 110.

<C. Functional configuration of analyzer>
FIG. 3 is a diagram showing the functional configuration of the analyzer. As shown in FIG. 3, the analysis device 100 includes an acquisition section 10, a storage section 11, a determination section 14, an estimation section 15, a correction section 16, and a suggestion information generation section 17. The acquisition unit 10, the determination unit 14, the estimation unit 15, the correction unit 16, and the proposal information generation unit 17 are realized by the processor 101 executing the program 110. Storage unit 11 is realized by ROM 102 and RAM 103.

The acquisition unit 10 acquires power amount data 12 for the analysis target period for the analysis target building 200 from the server device 300 via the network. The analysis period is, for example, one year. However, holidays (for example, Saturdays, Sundays, and holidays) of the building 200 to be analyzed are excluded from the analysis target period. The acquisition unit 10 stores the acquired power amount data 12 in the storage unit 11.

Further, the acquisition unit 10 acquires outside temperature data 13 indicating the outside temperature for each time period during the analysis target period in the area including the analysis target building 200 via the network. The outside temperature data is distributed from a predetermined organization (for example, the Japan Meteorological Agency). The acquisition unit 10 stores the acquired outside temperature data 13 in the storage unit 11.

The determination unit 14 uses the power amount data 12 to determine whether each day included in the analysis target period is an operating day or a non-operating day for the air conditioner (air conditioner).

The estimating unit 15 estimates the power consumption of the building 200 by use during the target time period of the target day included in the analysis target period. In this embodiment, the estimating unit 15 estimates power consumption for three uses. The power consumption for the three uses is base power consumption, activity power consumption, and air conditioning power consumption.

Base power consumption is power consumption for applications that are not due to human activity. Base power consumption includes the amount of power consumed by communication equipment, the amount of power consumed by cooling equipment that is constantly used throughout the year (e.g., air conditioners in server rooms, refrigerated freezer cases in food supermarkets), and the amount of power consumed by lighting outlets. Includes standby power consumption.

The activity power consumption is the power consumption for uses caused by human activities. Activity power consumption includes power consumed by ventilation power, power consumed by lighting outlets (excluding standby power), power consumed by elevators, and power used to charge EVs (Electric Vehicles). This includes the amount of electricity consumed by electric water heating equipment, and the amount of electricity consumed by other power sources.

The air conditioning power consumption is the power consumption for an application that adjusts temperature and/or humidity for the purpose of improving human comfort (hereinafter referred to as "air conditioning application"). The air conditioning power consumption includes the amount of power consumed by the heat source equipment and the amount of power consumed by the air conveying power of the air conditioner. Note that since ventilation is simply an adjustment of airflow, the amount of power consumed by ventilation power is not included in the air conditioning power consumption. The power consumption of cooling equipment that is constantly used throughout the year is included in the base power consumption as described above, and is not included in the air conditioning power consumption.

The correction unit 16 corrects the target power consumption when there is a difference between the target power consumption measured during the target time period on the target day and the sum of the base power consumption, activity power consumption, and air conditioning power consumption. At least one of the activity power consumption and the air conditioning power consumption is corrected so that the amount and the total match. The target power consumption amount is the power consumption amount indicated by the data corresponding to the target time period on the target date among the power amount data.

The proposal information generation unit 17 generates proposal information for the manager of the building 200 based on the base power consumption, the activity power consumption, and the air conditioning power consumption.

<D. Analysis processing flow>
FIG. 4 is a flowchart showing an example of the flow of analysis processing in the analyzer. As shown in FIG. 4, the acquisition unit 10 of the analysis device 100 first acquires the power amount data 12 and outside temperature data 13 for the analysis target period for the building 200 to be analyzed (step S1). The acquired power amount data 12 and outside temperature data 13 are stored in the storage unit 11.

Next, the estimation unit 15 estimates the base power consumption amount in the target time period on the target day using the power amount data 12 and the outside temperature data 13 (step S2).

Next, the determining unit 14 uses the power amount data 12 to determine whether each day included in the analysis target period is an operating day or a non-operating day for the air conditioner (step S3).

Next, the estimating unit 15 calculates the amount of electricity based on the data of the day determined as a non-working day by the determining unit 14 (hereinafter referred to as "non-working day data") out of the power amount data 12 and the base power consumption. Then, the power consumption amount for the activity in the target time period on the target day is estimated (step S4).

Next, the estimating unit 15 estimates the air conditioning power consumption in the target time period on the target day based on the non-working day data (step S5).

Next, the correction unit 16 corrects at least one of the air conditioning power consumption and the activity power consumption as necessary (step S6).

Next, the proposal information generation unit 17 generates proposal information for the manager of the building 200 based on the base power consumption, the activity power consumption, and the air conditioning power consumption (step S7).

Note that the processing order of steps S1 to S6 is not limited to the order shown in FIG. For example, step S3 may be performed before step S2. Further, step S5 may be executed before step S4.

<E. How to estimate base power consumption>
The base power consumption estimation method executed in step S2 will be described with reference to FIG. 5. FIG. 5 is a diagram illustrating a method for estimating base power consumption.

The estimation unit 15 uses the power consumption data 12 and the outside temperature data 13 to generate unit data that associates the power consumption of each time period on each day included in the analysis target period with the outside temperature of that time period. generate. For example, if the number of days included in the period to be analyzed is 300, and each day is divided into 96 time periods of 15 minutes each, the estimation unit 15 generates 300×96=28,800 unit data. The upper part of FIG. 5 shows the results of plotting a plurality of generated unit data 30 on a graph in which the horizontal axis is the outside temperature (° C.) and the vertical axis is the power consumption (kWh/15 min).

Base power consumption is power consumption for applications that are not due to human activity and is likely to correspond to the lowest power consumption at each outside temperature. Therefore, the estimation unit 15 estimates the base power consumption amount in the target time period on the target day according to the following steps (1) to (4).

Step (1): The estimation unit 15 divides the set outside temperature range into a plurality of outside temperature zones. The outside temperature range may be set in advance or may be set according to the distribution of the outside temperature data 13. For example, the range occupied by 95% of the outside temperature distribution shown by the outside temperature data 13 is determined as the outside temperature range. The width of the outside temperature range is predetermined, and is, for example, 1°C. In the middle part of FIG. 5, divided outside air temperature zones are shown. In the example shown in FIG. 5, the estimation unit 15 divides the outside air temperature range from 0 to 40°C into 40 outside air temperature zones each having a width of 1°C.

Step (2): The estimation unit 15 extracts the minimum power consumption for each outside temperature zone. The middle part of FIG. 5 shows an enlarged view of the graph shown in the upper part of FIG. 5 at 5 to 9 degrees Celsius. In the example shown in FIG. 5, for each of the 40 outside temperature zones, the estimation unit 15 extracts the unit data 31 having the minimum power consumption from the unit data 30 belonging to the outside temperature zone. The estimation unit 15 determines the power consumption amount indicated by the unit data 31 extracted for each of the 40 outside temperature zones as the minimum power consumption amount for that outside temperature zone.

Step (3): The estimation unit 15 calculates a regression equation using the outside temperature as an explanatory variable and the minimum power consumption extracted in step (2) as an objective variable. In the example shown in FIG. 5, the estimation unit 15 uses the outside temperature indicated by the unit data 31 extracted for each of the 40 outside temperature zones as an explanatory variable, and uses the power consumption (minimum power consumption) indicated by the unit data 31 as an explanatory variable. Calculate the regression equation using the objective variable (quantity) as the objective variable. The regression equation shows, for example, an n-dimensional curve and is calculated using the least squares method.

In the middle and lower rows of FIG. 5, a regression curve 33 expressed by a regression equation is shown. In the example shown in FIG. 5, the regression curve 33 shows a correlation with the outside temperature when the outside temperature is high. This is because the amount of power consumed by cooling equipment for communication equipment increases as the outside temperature rises.

Note that in an outside temperature zone in which the number of unit data 30 is small, there is a possibility that unit data 30 indicating only the power consumption amount for uses that are not caused by human activities is not included. Therefore, for an outside temperature zone in which the number of unit data 30 is less than the threshold, the estimation unit 15 uses the minimum power consumption of the outside temperature zone closest to the outside temperature zone in which the number of unit data 30 is greater than or equal to the threshold. It is preferable to calculate a regression equation using Note that the estimation unit 15 may calculate the regression equation for an outside temperature zone in which the number of unit data 30 is less than the threshold value using the intermediate temperature of the outside temperature zone as the outside temperature.

In the example shown in FIG. 5, since the number of unit data 30 belonging to the outside air temperature range of 5°C or lower is less than the threshold, the estimation unit 15 determines that the minimum power consumption of the outside air temperature range of 5°C or lower is 5 to 6. Calculate the regression equation using the minimum power consumption in the outside temperature range of °C. Furthermore, since the number of unit data 30 belonging to the outside air temperature range of 30°C or higher is less than the threshold, the estimation unit 15 determines that the outside air temperature range of 29 to 30°C is the minimum power consumption for the outside air temperature range of 30°C or higher. using the minimum power consumption of

Step (4): The estimation unit 15 estimates the base power consumption in the target time period on the target day based on the regression equation and the outside temperature in the target time zone on the target day. That is, the estimation unit 15 calculates the base power consumption amount in the target time period on the target day by substituting the outside temperature in the target time zone on the target day into the regression equation.

FIG. 6 is a diagram showing the verification results of the estimation accuracy of the base power consumption estimated by the estimator. FIG. 6(a) shows the verification results of the estimation accuracy of the base power consumption estimated using the winter season (January) as the target date. FIG. 6(b) shows the verification results of the estimation accuracy of the base power consumption estimated using the summer season (September) as the target date.

The verification was carried out by measuring the power consumption of each type of equipment installed in the building 200. FIG. 6 shows temporal changes in power consumption of each facility. FIG. 6 shows an example of a building 200 that includes only communication equipment cooling equipment (server room air conditioner) as cooling equipment used throughout the year. In FIG. 6, the amount of power indicated by "communication device system" includes the amount of power consumed by the communication device and the amount of power consumed by the cooling equipment for the communication device.

As shown in FIG. 6, the line 34 indicating the waveform of the base power consumption estimated by the estimation unit 15 is lower than the trough of the waveform indicating the sum of the power amount of the communication equipment system and the power amount of the light outlet. This shows a slightly smaller amount of electricity. The power amount of the light outlet includes the standby power amount of the light outlet and the amount of power when the light outlet is in use. In a waveform showing a temporal change in the amount of power of a light outlet, the valleys indicate the amount of power in standby mode, and the peaks indicate the amount of power in use. In this way, the base power consumption estimated by the estimator 15 is the power consumption of uses that are not caused by human activities (the amount of power consumed by communication equipment, the amount of power consumed by cooling equipment for communication equipment). , the standby power consumption of lighting outlets).

As shown in Figure 6, the estimated base power consumption in the winter (and intermediate seasons (spring and autumn) is approximately constant, whereas the estimated base power consumption in the summer season fluctuates slightly. This is because in the summer when the outside temperature is high, the amount of power consumed by cooling equipment for communication equipment increases as the outside temperature rises.

<F. How to determine the working days/non-working days of an air conditioner＞
Next, a method for determining whether the air conditioner is in operation or not in operation, which is executed in step S3, will be described. The determination unit 14 determines whether each day included in the analysis target period is an operating day or a non-operating day for the air conditioner, according to the following procedures (a) and (b).

Step (a): The determination unit 14 extracts, as training data, data on days in which the maximum power consumption among the power consumption measured for each time period of the day is lower than the predetermined number of days from the power amount data. do.

Specifically, for each day included in the analysis target period, the determination unit 14 identifies the maximum power consumption among the power consumption for each time period on that day. The determining unit 14 sorts the identified maximum power consumption amounts in ascending order, and defines the days of the lower predetermined number of days as days when the air conditioner is not in operation. The predetermined number of days is, for example, 20 days. The determination unit 14 extracts the data of the lower predetermined number of days from the power amount data, and generates training data by concatenating the extracted data.

Step (b): For each day included in the analysis target period, the determination unit 14 determines whether the air conditioner is determine whether it is a working or non-working day. Step (b) includes, for example, the following steps (b-1) to (b-4).

Step (b-1): The determination unit 14 creates test data that is time-series data in which the determination target data is combined after the training data.

Step (b-2): The determination unit 14 uses a change point detection algorithm to calculate a change point score at each time point in the test data. As a change point detection algorithm, for example, ChangeFinder ("J. Takeuchi and K. Yamanishi: A Unifying Framework for Detecting Outliers and Change Points from Time Series, IEEE Tran. on Knowledge and Data Engineering, Vol. 18, No. 4 (April- 2006), p.482-492" and "Kiyoyuki Kaito, Kosuke Kazumi: Early anomaly detection of social infrastructure facilities based on statistical change point detection (anomaly detection and change point detection), Reliability Society REAJ Journal, Vol. .37 (Vol. 223) (2015), p.116-125) can be used. ChangeFinder employs two-step learning of a time series model, and calculates a change point score after distinguishing the degree of change in time series data from outliers. Specifically, the model learned in the first stage is used to detect outliers, and the model learned in the second stage is used to detect changing points.

FIG. 7 is a diagram showing an example of application of ChangeFinder to test data. FIG. 7 shows a temporal change in test data in which judgment target data is combined after 20 days of training data.

The determination unit 14 performs first stage learning. The determination unit 14 uses an AR model (Auto Regressive Model) as a probability model of the test data, learns the AR model using an SDAR (Sequentially Discounting Auto Regressive) algorithm, and calculates a probability density function p _t (x) (t=1 ,2,...). The determination unit 14 uses the probability density function p _t (x) obtained through learning to calculate the logarithmic loss expressed by the following equation (1) as an outlier score of the data x _t at each time point t.
Score1(x _t )=-lnp _t-1 (x _t )...Formula (1)
In equation (1), Score1(x _t ) is an outlier score.

The determination unit 14 smoothes the outlier score. The determination unit 14 provides a window whose width is T, which is a positive integer, and calculates the average value of the outlier scores at the time points included in the window. The determination unit 14 calculates a time-series moving average score y _t (t=1, 2, . . . ) expressed by the following equation (2) by sliding the window.

Next, the determination unit 14 performs second stage learning. The determination unit 14 uses the AR model as a probability model for the new time series data y _t (t=1, 2, . . . ) obtained by equation (2), and learns the AR model using the SDAR algorithm. , calculate the probability density function q _t (y) (t=1, 2, . . . ). Further, the determination unit 14 determines the change point score at each time point t according to the following equation (3) using the logarithmic loss −lnq _t-1 (y _t ) at each time point t and T′, which is a positive integer. Calculate. In equation (3), Score2(y _t ) is the change point score.

It can be determined that the larger the Score2(y _t ), the greater the degree of change in the test data, which is time-series data, at time t.

Step (b-3): The determination unit 14 divides the change point score for each time point in the test data into a first data group at a time point corresponding to the training data and a second data group at a time point corresponding to the determination target data. To divide. Note that in the test data, the change point score becomes extremely large until a predetermined time elapses from the beginning because there is little learning data. Therefore, change point scores until a predetermined period of time has passed since the beginning (in the example shown in FIG. 7, five days from the beginning) are not included in the first data group.

Step (b-4): The determination unit 14 determines whether the change point score (first change point score) shown in the first data group and the change point score (second change point score) shown in the second data group are different. Determine whether they are equivalent or not by testing. The determination unit 14 uses, for example, Welch's t-test. The determination unit 14 determines that, as a result of the t-test, for example, when the p value is below the significance level of 5%, there is a significant difference between the average value of the first change point score and the average value of the second change point score ( are not equivalent). If there is a significant difference between the average value of the first change point score and the average value of the second change point score, the determination unit 14 determines that the day corresponding to the determination target data is an operating day of the air conditioner. do. If there is no significant difference between the average value of the first change point score and the average value of the second change point score, the determination unit 14 determines that the day corresponding to the determination target data is a day when the air conditioner is not in operation. judge.

The accuracy of the determination of working days/non-working days determined in this manner was verified. Specifically, when the determining unit 14 determined the working days/non-working days of the air conditioner using the electricity amount data of a building in which the actual number of working days of the air conditioner is 208 in one year, The 204th day was determined to be a working day. That is, the determination accuracy was 97.1%. In this way, it was confirmed that the determination accuracy by the determination unit 14 was sufficiently high.

<G. How to estimate activity power consumption>
The power consumption when the air conditioner is not operating is the sum of the base power consumption and the activity power consumption. Activity power consumption fluctuates over time. However, for small and medium-sized commercial buildings, the amount of electricity consumed by activities during a certain time period does not vary significantly throughout the year. Therefore, the estimating unit 15 estimates the amount of power consumed by activities in the target time period, assuming that the amount of power consumed by activities in each time period is constant throughout the year.

The estimating unit 15 calculates a first representative value (for example, an average value, an intermediate value, etc.) of power consumption in a target time period on a non-working day based on the non-working day data of the power amount data 12. Further, the estimation unit 15 estimates the base power consumption in the target time period on each non-working day according to the base power consumption estimation method described above, and estimates a second representative value (for example, the average value) of the estimated base power consumption. , intermediate values, etc.). The estimation unit 15 estimates the amount obtained by subtracting the second representative value from the first representative value as the activity power consumption amount in the target time period.

FIG. 8 is a diagram showing the verification results of the estimation accuracy of the activity amount power consumption estimated by the estimator. The verification was performed by measuring the power consumption of each type of equipment installed in the building 200, as in FIG. 6. FIG. 8 shows the verification results of the estimation accuracy of activity power consumption for each time period estimated from November 7th to November 11th. As described above, since the activity power consumption for each time period is constant throughout the year, the activity power consumption estimated for the same time period on the five target days is the same. Note that FIG. 8 also shows the base power consumption estimated for the target days from November 7th to November 11th.

As shown in FIG. 8, the activity power consumption estimated by the estimation unit 15 includes the amount of power consumed by ventilation equipment, the amount of power consumed by lighting outlets (excluding standby power), and the amount of power consumed by elevators. This approximately matches the sum of the actual measured values of the amount of power used for EV charging, the amount of power consumed by electric water heating equipment, and the amount of power consumed by other motive power. In this way, the activity-based power consumption estimated by the estimation unit 15 accurately represents the sum of the power consumption of uses caused by human activities.

<H. How to estimate power consumption for air conditioning＞
Referring to FIG. 9, the method for estimating the air conditioning power consumption amount executed in step S5 will be described. FIG. 9 is a diagram illustrating a method for estimating power consumption for air conditioning. The air conditioning power consumption corresponds to the power consumption that exceeds the power consumption when the air conditioner is not operating. Therefore, the estimating unit 15 estimates the air conditioning power consumption in the target time period on the target day as follows.

FIG. 9 shows the distribution of power consumption during the target time period from 14:00 to 14:15 in the analysis target period. In FIG. 9, circles indicate data on days determined by the determination section 14 as days when the air conditioner is in operation, and cross marks indicate data on days determined by the determination section 14 as non-operation days for the air conditioner. .

The estimating unit 15 calculates that <G. Method of estimating power consumption for activities> Similarly, based on the non-working day data of the power amount data 12, the first representative value (for example, average value, intermediate value, intermediate value, values, etc.). In FIG. 9, a line 35 indicates the average value of power consumption during the target time period on non-working days.

The estimation unit 15 calculates the amount obtained by subtracting the first representative value from the power consumption amount (hereinafter referred to as "target power consumption amount") indicated by the data 36 corresponding to the target time period on the target date among the power amount data. , estimated as the air conditioning power consumption during the target time period on the target day.

Note that if the target day is determined by the determination unit 14 to be a non-working day, the estimating unit 15 estimates the air conditioning power consumption in the target time zone of the target day to be 0.

FIG. 10 is a diagram showing the verification results of the estimation accuracy of the air conditioning power consumption estimated by the estimator. The verification was carried out by measuring the amount of power consumed by the air conveyance power of the heat source equipment and air conditioner installed in the building 200. FIG. 10 shows the verification results of the estimation accuracy of the air conditioning power consumption estimated for the target days from February 1st to February 9th. As shown in FIG. 10, the air conditioning power consumption estimated by the estimation unit 15 substantially matches the actual measurement value. In this way, the air conditioning power consumption estimated by the estimation unit 15 accurately represents the power consumed by the heat source equipment and the air conveyance power of the air conditioner.

<I. Correction method＞
Next, the correction method executed in step S6 will be explained. As described above, the base power consumption is estimated based on the regression equation and the outside temperature in the target time period on the target day. The power consumption for activities during the target time period is estimated using data on non-working days, and takes a constant value throughout the year. Therefore, the target power consumption indicated by the data corresponding to the target time period on the target date among the power consumption data, the sum of the base power consumption, activity power consumption, and air conditioning power consumption (hereinafter referred to as the "estimated value") There is a possibility that there will be a difference between the Therefore, the correction unit 16 corrects the activity power consumption and the air conditioning power consumption in the following manner.

An activity time period, which is a time period in which human activities are scheduled to be performed in the building 200, is predetermined in the analysis device 100. Specifically, the activity time period (for example, 8:00 to 20:00) for each building is registered in advance in the analysis device 100 using the operation keys 106 (see FIG. 2).

The main reason why the target power consumption becomes larger than the sum of the estimated values during active hours is thought to be mainly due to excessive operation of the air conditioner. Therefore, when the target time period is included in the activity time period and the target power consumption is larger than the sum of the estimated values, the correction unit 16 adjusts the air conditioning by the difference between the target power consumption and the sum of the estimated values. increase power consumption.

The first possible cause of the target power consumption being smaller than the sum of the estimated values during the activity period is the implementation of energy saving actions for air conditioners, and the second possible reason is the implementation of energy saving actions for other equipment. Therefore, the correction unit 16 determines that the target time period is included in the activity time period, the target power consumption is smaller than the sum of the estimated values, and the difference between the target power consumption and the sum of the estimated values is equal to the air conditioning amount. If it is less than or equal to the power consumption amount, the air conditioning power consumption is reduced by the difference. On the other hand, the correction unit 16 determines that the target time period is included in the activity time period, the target power consumption is smaller than the sum of the estimated values, and the difference between the target power consumption and the sum of the estimated values is equal to If the power consumption exceeds the power consumption by the excess amount, the power consumption for air conditioning is reduced to 0, and the power consumption for activities is reduced by the excess amount.

The main reason why the target power consumption becomes larger than the sum of estimated values during times other than active hours is thought to be mainly due to an increase in human activity. Therefore, when the target time period is not included in the activity time period and the target power consumption is larger than the sum of the estimated values, the correction unit 16 corrects the activity by the difference between the target power consumption and the sum of the estimated values. Increase power consumption.

The reasons why the target power consumption becomes smaller than the sum of the estimated values during times other than the active hours are firstly that the power consumption for air conditioning is overestimated, and secondly, the power consumption of other equipment It is possible to implement energy saving actions against Therefore, the correction unit 16 determines that the target time period is not included in the activity time period, the target power consumption is smaller than the sum of the estimated values, and the difference between the target power consumption and the sum of the estimated values is equal to the air conditioning amount. If it is less than or equal to the power consumption amount, the air conditioning power consumption is reduced by the difference. On the other hand, the correction unit 16 determines that the target time period is not included in the activity time period, the target power consumption is smaller than the sum of the estimated values, and the difference between the target power consumption and the sum of the estimated values is equal to the air conditioning amount. If the power consumption exceeds the power consumption by the excess amount, the power consumption for air conditioning is reduced to 0, and the power consumption for activities is reduced by the excess amount.

The correction by the correction unit 16 allows the measured actual power consumption to match the sum of the base power consumption, the activity power consumption, and the air conditioning power consumption.

<J. Verification results of power consumption decomposition accuracy>
By comparing the base power consumption, activity power consumption, and air conditioning power consumption estimated and corrected by the analyzer 100 with the actual values for five different buildings A to E, three uses of power consumption can be determined. We verified the decomposition accuracy.

FIG. 11 is a diagram showing the verification results of the accuracy of decomposition of power consumption into three uses. In FIG. 11, (a) to (e) show the results of comparing estimated values and actual values of base power consumption, activity power consumption, and air conditioning power consumption in buildings A to E, respectively. In addition, in FIGS. 11(a) to (e), the upper row shows the comparison results between the estimated value and the actual measured value of the base power consumption, and the middle row shows the estimated value and the actual measured value of the activity power consumption. The comparison result between the estimated value and the actual value of the air conditioning power consumption is shown in the lower row. Note that the estimated values of the base power consumption, the activity power consumption, and the air conditioning power consumption are appropriately corrected by the correction unit 16.

FIG. 12 is a diagram showing average values of errors between estimated values and actual measured values of base power consumption, activity power consumption, and air conditioning power consumption.

As shown in FIGS. 11 and 12, the errors between the estimated values and the actual measurements of each of the base power consumption, activity power consumption, and air conditioning power consumption are 15% or less. From this, the breakdown of power consumption into three uses is performed with high accuracy.

<K. Proposal information generation process>
The proposal information generation unit 17 generates, for example, the following first to third proposal information.

The first proposal information indicates the amount of energy saved (hereinafter referred to as the "first amount of energy saved") when an air conditioner that is operating at an outside temperature where the air conditioner does not need to be operated is stopped.

The second proposal information indicates the amount of energy savings (hereinafter referred to as "second amount of energy savings") that can be expected when the air conditioner, which is operating excessively with respect to the outside temperature, is changed to optimal operation.

The third proposal information indicates the amount of energy saved (hereinafter referred to as the "third amount of energy saved") that can be expected when it is prevented from forgetting to turn off the electric light outlet at night.

(K-1. First proposal information generation method)
The proposal information generation unit 17 extracts the outside temperature during the time period included in the period from 9:00 to 17:00 on the day when the air conditioner is determined to be non-operating from the outside temperature data. The proposal information generation unit 17 determines the lowest 25% of the outside temperatures indicated by the extracted outside temperature data as the heating start outside temperature, and determines the top 25% of the outside temperatures as the cooling start outside temperature. Determine as the starting outside temperature. The proposal information generation unit 17 determines the outside temperature range from the heating start outside temperature to the cooling start outside temperature as an air conditioning unnecessary temperature range where the air conditioner does not need to operate.

The proposal information generation unit 17 identifies a day determined to be an operating day of the air conditioner in the analysis target period as a target day, and further specifies a time period on the target day in which the outside temperature falls within the air conditioning unnecessary temperature range as the target time. Identify as a band. The proposal information generation unit 17 calculates the total value of the air conditioning power consumption estimated by the estimation unit 15 for the target time periods of these target days as the first energy saving amount. Then, the proposal information generation unit 17 generates first proposal information indicating the calculated first energy saving amount.

(K-2. Method of generating second proposal information)
The proposal information generation unit 17 determines the heating start outside temperature and the cooling start outside temperature using the same method as described above. The proposal information generation unit 17 extracts data on the estimated air conditioning power consumption amount as the heating outside temperature zone data, with the time period corresponding to the outside temperature below the heating start outside temperature as the target time period. The proposal information generation unit 17 extracts data on estimated air conditioning power consumption as cooling outside temperature zone data, using a time zone corresponding to an outside temperature equal to or higher than the cooling start outside temperature as the target time period.

The proposal information generation unit 17 applies robust regression to the heating outside temperature range data. Robust regression is a method that adds outlier testing to linear regression. By applying robust regression, a threshold value for determining outliers can be automatically set based on data variations. The proposal information generation unit 17 classifies the heating outside temperature range data into normal data and abnormal data using the set threshold. The proposal information generation unit 17 calculates the sum of the differences between the abnormal data and the threshold value as the second energy saving amount during heating operation.

Similarly, the proposal information generation unit 17 applies robust regression to the cooling outside temperature range data. The proposal information generation unit 17 classifies the cooling outside temperature zone data into normal data and abnormal data using a threshold set by robust regression. The proposal information generation unit 17 calculates the sum of the differences between the abnormal data and the threshold value as the second energy saving amount during cooling operation.

The proposal information generation unit 17 generates second proposal information indicating the calculated second energy saving amount during heating operation and cooling operation.

(K-3. Third proposal information generation method)
The proposal information generation unit 17 extracts data on power consumption for activities estimated by setting a time period other than the active time period as a target time period as nighttime activity data. The proposal information generation unit 17 applies robust regression to the nighttime activity data using the same method as (K-2. Second proposal information generation method). The proposal information generation unit 17 classifies the nighttime activity data into normal data and abnormal data using a threshold set by robust regression. The proposal information generation unit 17 calculates the sum of the differences between the abnormal data and the threshold value as the third energy saving amount. The proposal information generation unit 17 generates third proposal information indicating the calculated third energy saving amount.

<L. Modified example>
In the above description, it is assumed that the analysis device 100 is equipped with the proposal information generation section 17. However, the proposal information generation unit 17 may be included in an information processing device that is communicably connected to the analysis device 100. In this case, the information processing device obtains the base power consumption, the activity power consumption, and the air conditioning power consumption for each time period within the analysis target period from the analysis device 100, and uses the acquired data to generate the proposal information. All you have to do is generate .

When the proposal information generation unit 17 generates only one of the first proposal information and the second proposal information, the estimation unit 15 determines whether the base power consumption, the activity power consumption, and the air conditioning power consumption are based on the air conditioning power consumption. Only the amount of power consumption may be estimated. When the proposal information generation unit 17 generates only the third proposal information, the estimation unit 15 calculates the base power consumption and the activity power consumption out of the base power consumption, the activity power consumption, and the air conditioning power consumption. Only the amount may be estimated.

In the example shown in FIG. 1, smart meter 210, server device 300, and analysis device 100 exchange data with each other via a network. However, the method of exchanging data between smart meter 210, server device 300, and analysis device 100 is not limited to the example shown in FIG. 1. For example, server device 300 may directly acquire power amount data from smart meter 210 without going through a network. Further, the analysis device 100 may be connected to the server device 300 without going through a network, and may directly acquire power amount data of the building 200 from the server device 300.

<M. Summary＞
As described above, the analysis device 100 that analyzes the power consumption state of the building 200 includes the acquisition section 10, the determination section 14, and the estimation section 15. The acquisition unit 10 acquires power amount data 12 indicating the power consumption measured for each time period of the building 200 during the analysis target period. The determination unit 14 determines whether each day included in the analysis target period is an operating day or a non-operating day for the air conditioner. The estimating unit 15 estimates the power consumption by use of the building during the target time period on the target day included in the analysis target period based on non-working day data that is data on non-working days among the power amount data 12. . The determination unit 14 extracts, from the power amount data 12, data on days in which the maximum power consumption among the power consumption measured for each time period in a day is a lower predetermined number of days as training data. For each day included in the analysis target period, the determination unit 14 determines whether the day is a working day or a non-working day based on the comparison result between the training data and the determination target data that is the data of that day among the electric energy data 12. Decide which one.

According to the above configuration, the training data is data on days in which the maximum power consumption among the power consumption measured for each time period in one day is the lower predetermined number of days. Therefore, the training data corresponds to data on days when the air conditioner is likely not in operation. Therefore, based on the comparison result between the training data and the determination target data, it is possible to accurately determine each day included in the analysis target period as either an operating day or a non-operating day for the air conditioner. As a result, as much data as possible corresponding to days when the air conditioner is not in operation can be extracted from the power amount data 12 with high accuracy. Thereby, the power consumption of the air conditioner on non-operating days can be specified with high accuracy, and the power consumption of each purpose of the building 200 can be analyzed with high accuracy using the specified power consumption.

The embodiments disclosed this time should be considered to be illustrative in all respects and not restrictive. The scope of the present invention is indicated by the claims rather than the above description, and it is intended that all changes within the scope and meanings equivalent to the claims are included.

10 acquisition unit, 11 storage unit, 12 electric energy data, 13 outside temperature data, 14 determination unit, 15 estimation unit, 16 correction unit, 17 proposal information generation unit, 30, 31 unit data, 33 regression curve, 100 analysis device, 101 processor, 102 ROM, 103 RAM, 104 HDD, 105 communication IF, 106 operation keys, 107 power supply circuit, 108 display, 110 program, 200 building, 210 smart meter, 220 equipment, 300 server device, SYS system.

Claims (10)

An analysis device that analyzes the power consumption state of a building,
an acquisition unit that acquires power amount data indicating power consumption measured for each time zone of the building during the analysis target period;
a determination unit that determines whether each day included in the analysis target period is an operating day or a non-operating day of the air conditioner;
An estimating unit that estimates the power consumption of each use of the building during the target time period on the target day included in the analysis target period based on non-working day data that is data on the non-working days among the power amount data. and
The determination unit includes:
From the power amount data, data on days in which the maximum power consumption among the power consumption measured for each time period in one day is a lower predetermined number of days is extracted as training data;
For each day included in the analysis target period, based on the comparison result between the training data and the judgment target data that is the data for that day among the electricity amount data, it is determined whether the day is the working day or the non-working day. Analyzer to determine which one. The determination unit includes:
Create test data that is time series data that combines the judgment target data after the training data,
calculating a change point score at each time point in the test data using a change point detection algorithm;
It is determined by a test whether a first change point score at a time point corresponding to the training data among the test data and a second change point score at a time point corresponding to the determination target data among the test data are equivalent. , when the first change point score and the second change point score are equivalent, it is determined that the day corresponding to the determination target data is a non-working day, and the first change point score and the second change point score are equal. The analysis device according to claim 1, which determines that a day corresponding to the determination target data is a working day if the change point scores are not equivalent. The analyzer according to claim 2, wherein the test is a t-test. The estimation unit is
If the target day is determined to be the non-working day, the power consumption for air conditioning in the target time period on the target day is estimated to be 0,
When the target day is determined to be the working day, the target time on the non-working day is calculated from the target power consumption measured during the target time period on the target day and from the non-working day data. The analysis device according to any one of claims 1 to 3, wherein an amount obtained by subtracting a representative value of power consumption of a zone is estimated as the air conditioning power consumption in the target time period of the target day. The acquisition unit further acquires outside temperature data indicating the outside temperature for each time zone in the area including the building during the analysis target period,
The estimation unit is
Extracting the minimum power consumption for each outside temperature zone using the power amount data and the outside temperature data,
Calculate a regression equation with outside temperature as the explanatory variable and the extracted minimum power consumption as the objective variable,
The analysis device according to claim 4, wherein the base power consumption amount in the target time period on the target day is estimated based on the regression equation and the outside temperature in the target time zone on the target day. The estimation unit is
The amount obtained by subtracting the representative value of the base power consumption in the target time period on the non-working day from the representative value of the power consumption in the target time period on the non-working day, which is calculated from the non-working day data. The analysis device according to claim 5, wherein is estimated as the activity-based power consumption amount of a use caused by human activity in the target time period. If a difference occurs between the target power consumption and the sum of the air conditioning power consumption, the base power consumption, and the activity power consumption, the target power consumption and the sum match. 7. The analysis device according to claim 6, further comprising a correction unit that corrects at least one of the air conditioning power consumption and the activity power consumption. An activity time period is determined in advance, which is a time period during which human activities are scheduled to take place in the building;
The correction unit is
If the target time period is included in the activity time period and the target power consumption is larger than the total sum, increase the air conditioning power consumption by the difference;
If the target time period is included in the activity time period, the target power consumption is smaller than the total sum, and the difference is less than or equal to the air conditioning power consumption, the air conditioning power consumption Reduce by the above difference,
If the target time period is included in the activity time period, the target power consumption is smaller than the total sum, and the difference exceeds the air conditioning power consumption by an excess amount, the air conditioning power consumption reducing the power consumption to 0 and reducing the power consumption of the activity by the excess amount;
If the target time period is not included in the activity time period and the target power consumption is larger than the total sum, increase the activity power consumption by the difference;
If the target time period is not included in the activity time period, the target power consumption is smaller than the total, and the difference is less than or equal to the air conditioning power consumption, the air conditioning power consumption Reduce by the above difference,
If the target time period is not included in the activity time period, the target power consumption is smaller than the total sum, and the difference exceeds the air conditioning power consumption by an excess amount, the air conditioning power consumption The analysis device according to claim 7, wherein the power consumption is reduced to 0 and the power consumption for the activity is reduced by the excess amount. An estimation method for estimating the power consumption state of a building,
a step in which the processor acquires power amount data indicating power consumption measured for each time period of the building in the analysis target period;
a step in which the processor determines whether each day included in the analysis target period is an air conditioning working day or a non-working day;
The processor calculates the power consumption by use of the building in the target time period of the target day included in the analysis target period based on the non-working day data that is the data of the non-working day among the power amount data. and a step of estimating,
The determining step includes:
a step in which the processor extracts, from the power amount data, data on days in which the maximum power consumption among the power consumption measured for each time period in one day is a lower predetermined number of days as training data;
For each day included in the analysis target period, the processor determines whether the day is the operating day or the and determining whether it is a non-working day. A program that causes a computer to execute an estimation method for estimating the power consumption state of a building,
The estimation method is
acquiring power amount data indicating power consumption measured for each time period of the building during the analysis target period;
determining whether each day included in the analysis target period is an air conditioning working day or a non-working day;
estimating power consumption by use of the building during a target time period on a target day included in the analysis target period based on non-working day data that is data on the non-working days among the power amount data; Equipped with
The determining step includes:
extracting, from the power amount data, data on days in which the maximum power consumption among the power consumption measured for each of the time periods in one day is a lower predetermined number of days as training data;
For each day included in the analysis target period, based on the comparison result between the training data and the judgment target data that is the data for that day among the electricity amount data, it is determined whether the day is the working day or the non-working day. and determining whether the program is a program.

Images