JP2013162589A - Calculation device, calculation method, and calculation program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To facilitate an estimation processing operation.SOLUTION: A calculation device 101 can access a storage area that stores the fluctuation amount of electric energy of a distribution board 1, a distribution board 2, and an electric outlet 1. Under this condition, with respect to an estimated content 103 estimating the connection between the distribution board 1 and the outlet 1, the calculation device 101 calculates a likelihood that can be a probability indicating the correctness-degree of the estimated content 103. The higher a likelihood is, the more the correctness of an assumption is proved. In the same manner, with respect to an estimated content 104 estimating the connection between the distribution board 2 and the outlet 1, the calculation device 101 calculates a likelihood that can be a probability indicating the correctness-degree of the estimated content 104.

Description

  The present invention relates to a calculation device, a calculation method, and a calculation program.

  In recent years, a technique has been disclosed in which a device connected between a power outlet and a device measures the amount of power. In addition, a technique for facilitating power management using such a technique is disclosed. For example, refer to the change in the power consumption of the distribution board when the adapter that controls the power of the outlet sequentially turns the power supply of the outlet from off to on in accordance with the instructions of the power management device. Thus, there is a technique for estimating the connection relationship. Also, by storing the general power consumption for each home appliance and comparing the power consumption measured with the distribution board with the stored power consumption, the connection relation between the distribution board and the home appliance There exists a technique for estimating (see, for example, Patent Documents 1 and 2 below).

JP 2010-213411 A JP 2010-193638 A

  However, in the above-described prior art, in order to estimate the connection relationship between the distribution board and the outlet, the adapter controls the power supply of the target device, or the building manager in advance It will prepare electric energy. Therefore, it is difficult for the manager of the building to instruct the execution request of the estimation process to the apparatus that executes the connection relation estimation process. For example, when controlling the power supply of the target device, the adapter turns on / off the power supply of the target device. Therefore, an apparatus that executes the estimation process cannot execute the estimation process unless the target device is not used by the user. In addition, regarding the preparation of general power consumption for each home appliance, the building manager specifies what home appliance is being used at the supply destination. Therefore, it takes time to prepare for execution of the estimation process, and it is difficult for the building manager to give an instruction to execute the estimation process to the apparatus that executes the estimation process.

  An object of the present invention is to provide a calculation device, a calculation method, and a calculation program capable of easily executing an estimation process in order to solve the above-described problems caused by the prior art.

  In order to solve the above-described problem and achieve the object, according to one aspect of the present invention, the amount of change of the measurement value of the measurement target at each supply source of the supply source group supplying the measurement target and the time The first correspondence information that is time information indicating the measurement time for each predetermined time, the amount of change of the measurement value of the measurement target at the supply destination to which the measurement target is supplied, and the measurement time for the predetermined time Supply from any of the supply source groups in which the time information of the first correspondence information and the time information of the second correspondence information are the same time from the storage unit that stores the second correspondence information serving as the time information to be shown Two or more combinations of the original change amount and the change amount at the supply destination at the same time are extracted, and two or more combinations of the extracted change amount at any of the supply sources and the change amount at the supply destination Based on which one of the suppliers and destinations is connected Possibility calculation device for calculating an index value indicating the calculation method and calculation program is proposed.

  According to one aspect of the present invention, it is possible to facilitate the execution of the estimation process.

FIG. 1 is an explanatory diagram illustrating an operation example of the calculation apparatus. FIG. 2 is an explanatory diagram showing a connection example of the power sensor network. FIG. 3 is an explanatory diagram illustrating a connection example of the communication system. FIG. 4 is an explanatory diagram illustrating a hardware example of the communication system. FIG. 5 is a block diagram illustrating an example of functions of the calculation apparatus. FIG. 6 is a sequence diagram illustrating an operation example of the communication system. FIG. 7 is an explanatory diagram showing an example of the contents stored in the measured power amount database. FIG. 8 is an explanatory diagram of an example of the contents stored in the power amount change amount table. FIG. 9A is an explanatory diagram (part 1) illustrating an operation example of the method 1 in the connection relationship estimation unit. FIG. 9B is an explanatory diagram (part 2) of an operation example of the method 1 in the connection relationship estimation unit. FIG. 10A is an explanatory diagram (part 1) illustrating an operation example of the method 2 in the connection relationship estimation unit. FIG. 10B is an explanatory diagram (part 2) of an operation example of the method 2 in the connection relationship estimation unit. FIG. 11A is an explanatory diagram (part 1-1) illustrating an operation example of the method 3 in the connection relationship estimation unit. FIG. 11B is an explanatory diagram (part 1-2) of an operation example of the method 3 in the connection relationship estimation unit. FIG. 12 is an explanatory diagram (part 2) of an operation example of the method 3 in the connection relationship estimation unit. FIG. 13 is an explanatory diagram (part 3) of an operation example of the method 3 in the connection relationship estimation unit. FIG. 14A is an explanatory diagram (part 1) illustrating an output example in the result output unit. FIG. 14B is an explanatory diagram (part 2) illustrating an output example in the result output unit. FIG. 15 is a flowchart illustrating an example of the method 1 in the connection relationship estimation processing procedure. FIG. 16A is a flowchart (part 1) illustrating an example of method 2 in the connection relationship estimation processing procedure. FIG. 16B is a flowchart (part 2) illustrating an example of method 2 in the connection relationship estimation processing procedure. FIG. 17A is a flowchart (part 1) illustrating an example of method 3 in the connection relationship estimation processing procedure. FIG. 17B is a flowchart (part 2) illustrating an example of method 3 in the connection relationship estimation processing procedure. FIG. 18 is a flowchart illustrating an example of a mixed normal distribution parameter setting processing procedure.

  Exemplary embodiments of a calculation device, a calculation method, and a calculation program according to the present invention will be described below in detail with reference to the accompanying drawings.

  FIG. 1 is an explanatory diagram illustrating an operation example of the calculation apparatus. The power sensor network 100 uses the distribution board 1 and the distribution board 2 for supplying the power supplied from the power company to the equipment in the building, and the distribution panel 1 or the distribution board 2 as a power source. Including an outlet 1 as a supply destination. Here, the building manager is in an unknown state whether the outlet 1 is connected to the distribution board 1 or the distribution board 2. In FIG. 1, a state in which the connection relation between the distribution board 1, the distribution board 2 and the outlet 1 is unknown is represented by “connection relation?”. Further, the calculation device 101 collects the power amount of the distribution board 1, distribution board 2, and outlet 1 via the communication network, and distributes the distribution board 1, distribution board 2, and outlet 1. The amount of power change is stored. The power amount change amount is a change amount of the power amount for each time. In the present embodiment, the calculation apparatus 101 estimates which supply source the outlet 1 is connected to from the amount of change in the electric energy.

The amount of power change stored in the calculation device 101 is, for example, the content shown by the graph 102. The horizontal axis of the graph 102 indicates time. The vertical axis of the graph 102 indicates the amount of power change. Graph 102 illustrates the amount of power variation of the outlet 1 and the distribution board 1 and the distribution board 2 with respect to time t ₁ ~ time t _3. With reference to the power amount change amount, the calculation apparatus 101 estimates the connection relationship between the distribution board 1 and the outlet 1 and the connection relationship between the distribution board 2 and the outlet 1. Specifically, the calculation device 101 calculates a likelihood that is a probability indicating the degree of correctness of the estimated content 103 with respect to the estimated content 103 in which the connection between the distribution board 1 and the outlet 1 is estimated. Similarly, the calculation device 101 calculates a likelihood that is a probability indicating the degree of correctness of the estimated content 104 with respect to the estimated content 104 in which the connection between the distribution board 2 and the outlet 1 is estimated.

  When the supply source and the supply destination are connected, if the power amount of the supply destination increases, the power amount of the supply source is likely to increase. In the present embodiment, the connection relationship is estimated using a feature that, when the power amount of the supply destination increases, the possibility that the power amount of the supply source increases increases. Further, it is assumed that the distribution board 1 and the outlet 1 are actually connected.

  First, the calculation device 101 calculates the likelihood of the estimated content 103. Specifically, the calculation device 101 calculates the probability that the amount of change in the electric energy of the distribution board 1 when the connection between the distribution board 1 and the outlet 1 is assumed at each time. Here, the probability that the amount of power change of the distribution board 1 can take means the probability that a certain amount of power change occurs in the group of power amount changes that can be generated in the distribution board 1. Next, the calculation device 101 calculates the likelihood of the estimated content 103 based on the probability at each time. For example, the calculation device 101 calculates the likelihood of the estimated content 103 by multiplying the probabilities at the respective times.

Specifically, in FIG. 1, the calculation device 101 assumes that the amount of power change that can occur in the distribution board 1 follows a normal distribution. Under this assumption, the calculation apparatus 101 generates a normal distribution from the average and standard deviation of the amount of power change of the distribution board 1, and uses the probability density function of the generated normal distribution as the amount of power change of the distribution board 1. Is a probability density function of the probability that can be taken. Next, the calculation apparatus 101 calculates the likelihood of the estimated content 103 using the probability density function f (x). Note that x represents the amount of power change. The probability density function f _C1 (x) of the amount of power change of the distribution board 1 assuming connection is the amount of power change obtained by subtracting the amount of power change of the outlet 1 from the amount of power change of the distribution board 1. Can be approximated by a probability density function of the probability that can be taken.

A graph 105 shows the probability density function f (x) of the generated normal distribution and the probability density function f _C1 (x) at time t1 in the estimation content 103. A graph 105 shows a probability density function that can be taken by the amount of change in the amount of power in the distribution board 1 at time t ₁ . The horizontal axis of the graph 105 indicates the amount of power change. The vertical axis of the graph 105 indicates the occurrence probability of the power amount change amount. f (x) is indicated by a broken line in the graph 105. f _C1 (x) is indicated by a solid line in the graph 105. C1 (t ₁ ) indicates the amount of change in electric energy at the outlet 1 at time t ₁ . B1 (t ₁ ) indicates the amount of change in the electric energy in the distribution board 1 at time t ₁ .

From the graph 102, C1 (t ₁ ) = 10 [W] and B1 (t ₁ ) = 15 [W]. Further, f _C1 (x) can be approximated as f _C1 (x) = f (x−10). Therefore, f _C1 (x) is a function moved in the positive direction from f (x). The probability f _C1 (15) that can be taken by the amount of power change 15 [W] in the distribution board 1 is larger than f (15). Thus, when the distribution board 1 and the outlet 1 are connected, if the amount of power change at the outlet 1 is positive, the amount of power change at the distribution board 1 is also likely to be positive. The probability that the amount of power change at 1 can be taken is also a large value.

Subsequently, the calculation unit 101, time t _2, the time t _3, ..., the amount of power variation in the distribution board 1 also time t _k is calculated the probability of possible. k indicates a time index. Table 106 shows an example of the calculated results. For time t _1, it is positive are _{C1 (t 1), B1 (} t 1) for the positive, the probability f _C1 at time t ₁ (t ₁₎ is larger than f (t _1). The probability at the time t ₂ and the probability at the time t ₃ are also large values by performing the same calculation method. Subsequently, the calculation apparatus 101 calculates the likelihood of the estimation content 103 by multiplying the probability at time t ₁ , the probability at time t ₂ , the probability at time t ₃ ,..., The probability at time t _k. .

  Next, the calculation device 101 calculates the likelihood of the estimated content 104. Specifically, the calculation device 101 calculates the probability that the amount of change in the electric energy of the distribution board 2 can be assumed when connection between the distribution board 2 and the outlet 1 is assumed at each time. Next, the calculation apparatus 101 calculates the likelihood of the estimated content 104 from the probability at each time. For example, the calculation device 101 calculates the likelihood of the estimated content 104 by multiplying the probabilities at each time. As described above, the calculation apparatus 101 calculates the likelihood of the estimated content 104 by a method similar to the method of calculating the likelihood of the estimated content 103.

A graph 107 shows the probability density function f (x) and the probability density function f _C1 (x) of the generated normal distribution in the estimation content 104. A graph 107 shows a probability density function that can be taken by the amount of power change in the distribution board 2 at time t ₁ . The horizontal axis of the graph 107 indicates the amount of power change. The vertical axis of the graph 107 indicates the occurrence probability of the amount of power change. f (x) is indicated by a broken line in the graph 107. f _C1 (x) is indicated by a solid line in the graph 107. B2 (t ₁ ) indicates the amount of change in the electric energy at the distribution board 2 at time t ₁ .

From the graph 102, C1 (t ₁ ) = 10 [W] and B2 (t ₁ ) = − 5 [W]. Further, f _C1 (x) can be approximated as f _C1 (x) = f (x−10). Therefore, f _C1 (x) is a function moved in the positive direction from f (x). From the graph 107, the probability f _C1 (−5) that can be taken by the power amount change amount −5 [W] in the distribution board 2 is smaller than f (−5).

Subsequently, the calculation unit 101, time t _2, the time t _3, ..., the amount of power variation in the distribution board 2 also time t _k is calculated the probability of possible. Table 108 shows an example of the calculated result. At time t ₁ , C 1 (t ₁ ) is positive and B 2 (t ₁ ) is negative, so the probability f _C1 (t ₁ ) at time t ₁ is smaller than f (t ₁ ). Subsequently, the time t _2, is positive are _{C1 (t 2), B2 (} t 2) because is positive, the probability f _C1 (t ₂₎ at time t ₂ is greater than f (t ₂₎ Become.

Furthermore, since the probability at time t ₃ is positive for C 1 (t ₃ ) and negative for B 2 (t ₃ ), the probability f _C1 (t ₃ ) at time t ₃ is f (t ₃ ). Smaller. Subsequently, the calculation apparatus 101 calculates the likelihood of the estimated content 104 by multiplying the probability at time t ₁ , the probability at time t ₂ , the probability at time t ₃ ,..., The probability at time t _k. .

When the likelihood of the estimated content 103 and the likelihood of the estimated content 104 are compared, the likelihood of the estimated content 103 having a large number of fields whose probability f _C1 (t _k ) is “large” is shown in Tables 106 and 108. Becomes a value higher than the likelihood of the estimated content 104.

  As described above, the calculation device 101 obtains the likelihood when connection or disconnection is assumed. Thereby, the calculation apparatus 101 can estimate the connection relation between the distribution board and the outlet without following the wiring. In addition, the calculation device 101 can easily execute the estimation process. Hereinafter, the details of the calculation device 101 will be described with reference to FIGS.

  FIG. 2 is an explanatory diagram showing a connection example of the power sensor network. The power sensor network 100 includes a distribution board 1, a distribution board 2, a parent distribution board 3, an outlet 1, a PC (Personal Computer) 201, an outlet 2, a display 202, a PC 203, and an outlet 3. , Facsimile 204 and printer 205. Among the devices included in the power sensor network 100, the distribution board 1, the distribution board 2, and the parent distribution board 3 are supply sources that supply electric power. In addition, among the devices included in the power sensor network 100, devices other than the distribution board 1, the distribution board 2, and the parent distribution board 3 are supply destinations to which power is supplied. A supply source exists upstream of the power sensor network 100 and a supply destination exists downstream of the power sensor network 100. Although not shown, it is assumed that there are N distribution boards in the power sensor network 100. N is an integer of 2 or more.

  Distribution board 1, distribution board 2, and parent distribution board 3 include power sensor B1, power sensor B2, and power sensor B3, respectively. Outlet 1 to outlet 3 include power sensor C1 to power sensor C3, respectively. The power sensors B1 to C3 are devices that measure power.

  The PC 201 and the PC 203 are computers that are occupied and used by individuals. The display 202 is a device that displays video. The display 202 displays the video output from the PC 201, for example. The PC 201 is connected to the outlet 1. Therefore, the amount of power used by the PC 201 is measured by the power sensor C1. Similarly, the display 202 is connected to the outlet 2. Therefore, the amount of power used by the PC 201 is measured by the power sensor C2.

  The facsimile 204 is a device that transmits image information to a remote place through a communication line. The facsimile 204 is connected to the outlet 3, and the power used by the facsimile 204 is measured by the power sensor C3. The printer 205 is a device that outputs information from a computer. For example, the printer 205 outputs information from the PC 203.

  The parent distribution board 3 supplies power to the distribution board 1 and the distribution board 2. The distribution board 1 and the distribution board 2 supply power to any device in the device group that is the supply destination. The user does not know the connection relationship for which device to supply specifically, and is unknown. In FIG. 2, as in FIG. 1, a state in which the connection relationship between distribution board 1, distribution board 2 and any device in the device group that is the supply destination is unknown is represented by “connection relationship?”. . The reason why the connection relationship becomes unclear is, for example, a case where pipe wiring work is performed due to aging or expansion of the building, and the connection relationship is different from that at the time of building design. Further, there are cases where the connection relationship should be grasped, such as when renovating a building or when buying or selling a building. At this time, it takes time and effort for the building manager to grasp the state of the building and the infrastructure. In the present embodiment, the calculation device 101 estimates the connection relationship for a state in which the connection relationship between the supply source and the supply destination is unknown.

  FIG. 3 is an explanatory diagram illustrating a connection example of the communication system. The communication system 300 includes a power sensor B1 to a power sensor B3, a power sensor C1 to a power sensor C3, a router 301, a hub 302, a wireless hub 303, and a calculation device 101. The communication system 300 includes a network 304 and a network 305. The network 304 includes a router 301, a hub 302, and a wireless hub 303. The network 305 includes a router 301, a calculation device 101, and a power sensor C3.

  The router 301 is a communication device that interconnects the network 304 and the network 305. The hub 302 is a device that relays electric transmission signals. Specifically, the hub 302 relays a power transmission signal performed by a device connected to the network 304 and the power sensors B1 to B3. The wireless hub 303 is a device that relays a transmission signal by wireless communication. Specifically, the wireless hub 303 relays a transmission signal performed by a device connected to the network 304, the power sensor C1, and the power sensor C2 through wireless communication.

  A network 304 and a network 305 connect a plurality of computers. The network 304 and the network 305 are, for example, a LAN (Local Area Network) or a WAN (Wide Area Network).

  FIG. 4 is an explanatory diagram illustrating a hardware example of the communication system. 4, the power sensor C3 among the power sensors B1 to C3 and the hardware of the calculation device 101 will be described. In FIG. 4, the power sensor C <b> 3 includes a CPU (Central Processing Unit) 401, a ROM (Read-Only Memory) 402, and a RAM (Random Access Memory) 403. The power sensor C3 includes a flash ROM 404, a sensor 405, and an IF (InterFace) 406. Each unit is connected by a bus 407.

  Here, the CPU 401 is an arithmetic processing device that controls the entire power sensor C3. The ROM 402 is a non-volatile memory that stores a program such as a boot program. A RAM 403 is a volatile memory used as a work area for the CPU 401. The flash ROM 404 is a nonvolatile memory that can rewrite stored contents. For example, the flash ROM 404 is a NOR type flash memory. Specifically, the flash ROM 404 stores firmware or the like that controls the operation of the power sensor C3. For example, when updating firmware, the power sensor C3 receives new firmware via the IF 406. Next, the power sensor C3 updates the old firmware stored in the flash ROM 404 to the received new firmware.

  The IF 406 is connected to the network 305 through a communication line, and is connected to other devices via the network 305. The IF 406 controls an internal interface with the network 305 and controls data input / output from an external device. The sensor 405 is a device that measures a measurement target.

  The calculation device 101 includes a CPU 411, a ROM 412, a RAM 413, a magnetic disk drive 414, a magnetic disk 415, and an IF 416. Each unit is connected by a bus 417.

  Here, the CPU 411 is an arithmetic processing device that controls the entire calculation device 101. The ROM 412 is a nonvolatile memory that stores programs such as a boot program. The RAM 413 is a volatile memory used as a work area for the CPU 411. The magnetic disk drive 414 is a control device that controls reading / writing of data with respect to the magnetic disk 415 according to the control of the CPU 411. The magnetic disk 415 is a non-volatile memory that stores data written under the control of the magnetic disk drive 414.

  The IF 416 is connected to the network 305 through a communication line, and is connected to other devices via the network 305. The IF 416 serves as an internal interface with the network 305 and controls data input / output from an external device. For example, a modem or a LAN adapter can be employed as the IF 416. The power sensor C3 and the calculation device 101 are connected via IF 406 and IF 416. Note that the calculation device 101 may include an optical disc drive, a display, a keyboard, and a mouse.

(Function of the calculation device 101)
Next, functions of the calculation apparatus 101 will be described. FIG. 5 is a block diagram illustrating an example of functions of the calculation apparatus. The calculation apparatus 101 includes a measurement data reception unit 501, a measurement data recording unit 502, a change amount generation unit 503, a connection relationship estimation unit 504, an index value recording unit 505, and a result output unit 506. Further, the calculation apparatus 101 can access the measured power amount database 507 and the power amount change amount table 508.

  The change amount generation unit 503 includes a measurement data reading unit 511, a supply source extraction unit 512, a power amount change amount calculation unit 513, a measurement sampling rate adjustment unit 514, and a change amount recording unit 515. Furthermore, the connection relationship estimation unit 504 includes an extraction unit 521, a generation unit 522, a setting unit 523, a calculation unit 524, a calculation unit 525, and an output unit 526. The measurement data receiving unit 501 to the output unit 526 serving as the control unit realize their functions when the CPU 411 executes the program stored in the storage device. Specifically, the storage device is, for example, the ROM 412, the RAM 413, the magnetic disk 415, or the like shown in FIG. Alternatively, the function may be realized by being executed by another CPU via the IF 416.

  The measurement data receiving unit 501 receives the amount of power from the power sensors B1 to C3. For example, the measurement data receiving unit 501 periodically receives the amount of power from the power sensor B1. Further, the measurement object to be measured may be a fluid supplied by piping such as gas and water in addition to the electric power described in the present embodiment. The received measurement value is stored in a storage area such as the RAM 413.

  The measurement data recording unit 502 records the power amount received by the measurement data reception unit 501 in the measurement power amount database 507 in association with the transmission source device and the time. As a method of associating with the transmission source device, for example, the power amount is associated with the identification information of the transmission source device. The identification information of the transmission source device may be a name such as “distribution panel 1” or “outlet 1”, or may be information unique to the power sensor. For example, the measurement data receiving unit 501 is 1740.0 [Wh] as the amount of power supplied by the distribution board 1 to “2011/9/6 12:00” from the power sensor B1 included in the distribution board 1 Assume that the amount of power is received. At this time, the measurement data recording unit 502 measures the power amount of 1740.0 [Wh] in association with the identification information of the distribution board 1 and the received time information “2011/9/6 12:00”. Record in the electric energy database 507.

  The change amount generation unit 503 generates a power amount change amount from the power amount. For example, the measured power amount database 507 indicates that the power amount in the distribution board 1 is “2011/9/6 12:00” and 1740.0 [Wh], and “2011/9/6 12:10”. ”Is stored as an electric energy of 1732.0 [Wh]. At this time, the change amount generation unit 503 generates a power amount change amount of 1732.0-1740.0 = −8 [Wh / 10 minutes] = − 48 [W]. The change amount generation unit 503 records the generated power amount change amount in the power amount change amount table 508.

  The connection relationship estimation unit 504 estimates the correspondence between the supply source and the supply destination from the power amount change amount table 508. For example, the connection relationship estimation unit 504 outputs an index value indicating that there is a high possibility that the outlet 1 is connected to the distribution board 1 from the power amount change amount table 508. In this embodiment, a specific estimation method is exemplified from method 1 to method 3. Method 1 will be described later with reference to FIGS. 9A and 9B. Method 2 will be described later with reference to FIGS. 10A and 10B. Method 3 will be described later with reference to FIGS. 11A to 13.

  The index value recording unit 505 records the index value that is output by the connection relationship estimation unit 504 and indicates the possibility that the supply source and the supply destination are connected. For example, a magnetic disk 415 or the like is employed as the recording destination.

  The result output unit 506 outputs the recorded index value. For example, a message that “the distribution board that may be connected to the outlet 1 is the distribution board 1” is output to the apparatus that remotely operates the calculation apparatus 101. As an output destination, the result output unit 506 may output the index value to the RAM 413 or may output the index value to a printer or display via the network 305.

  The measured power amount database 507 stores the power amount associated with the transmission source device and the time. For example, the measured power amount database 507 stores the power amount 1740.0 [Wh] associated with the time information “2011/9/6 12:00” in association with the transmission source. Details of the stored contents of the measured power amount database 507 will be described later with reference to FIG.

  The power amount change amount table 508 is the first time information indicating the amount of change per predetermined time of the measurement value of the measurement target at each supply source of the supply source group that supplies the measurement target and the measurement time per predetermined time. Store correspondence information. Furthermore, the power amount change amount table 508 is second correspondence information that is time information indicating a change amount per predetermined time and a measurement time per predetermined time of the measurement value of the measurement target at the supply destination to which the measurement target is supplied. Remember. For example, as an example of the first correspondence information, the amount of power change in distribution board 1 is −48 [W] and the measurement time is “2011/9/6 12:10”. Further, as an example of the second correspondence information, the amount of power change at the outlet 1 is -10 [W] and the measurement time “2011/9/6 12:10”. Details of the first correspondence information and the second correspondence information will be described later with reference to FIG.

  The predetermined time may be any time. In the present embodiment, the predetermined time is 10 minutes. Further, the predetermined time is preferably such that the measurement values of the measurement objects at the supply source and the supply destination become independent every predetermined time. If the predetermined time is too short, there is a possibility that the change amount at the first predetermined time and the change amount at the next predetermined time are not independent. For example, suppose that a certain home appliance shows an operation of transitioning from a first state in which power consumption gradually increases when the switch is turned on to a second state in which power consumption is constant. Further, the power amount change amount table 508 stores the first power amount change amount near the state where the power consumption amount increases, and before the state where the power consumption amount increases ends, the power amount It is assumed that the change amount table 508 stores the second power amount change amount.

  At this time, the second power amount change amount is related to the first power amount change amount. Therefore, the calculation device 101 may not be able to calculate a correct estimation result because the assumption that the amount of power change at each time is independent, which is assumed on the basis of the normal distribution and χ square distribution used in the present embodiment, cannot be observed. is there. Therefore, the predetermined time is preferably equal to or longer than a certain time interval.

  The time information indicating the measurement time for each predetermined time is the time when the measurement amount within the predetermined time is measured. Specifically, the time information may be both a start time and an end time of a predetermined time, or may be either one or may store a time that is the center of the start time and the end time.

  For example, in the power amount change amount table 508, the change amount in the distribution board 1 at “2011/9/6 12:10” is −48 [W], and the change is “2011/9/6 12:20”. Record the amount as -16 [W]. Details of the contents stored in the power amount change amount table 508 will be described later with reference to FIG.

  The measurement data reading unit 511 reads power sensor identification information, power amount, and time information from the measured power amount database 507. The timing for performing the reading is, for example, when a request for estimating the correspondence between the supply source and the supply destination is received from a device operated by the user. The estimation request is a request for estimating whether or not there is a high possibility of being connected to a specific supply destination among a plurality of supply sources. Furthermore, by making a plurality of estimation requests from a device operated by the user, the calculation device 101 can estimate connection sources of a plurality of supply destinations. Each read information is stored in a storage area such as the RAM 413.

  The supply source extraction unit 512 extracts a supply source that can be an estimation target from the supply source group. Specifically, the supply source extraction unit 512 sets, as a candidate for estimation target, a distribution board that supplies an amount of power equal to or higher than the power amount of the supply destination for which the power amount of the distribution panel is to be estimated at a certain time. It is obvious that a distribution board that supplies an amount of power less than the amount of power of the supply destination to be estimated is not connected to the supply destination to be estimated. Therefore, the supply source extraction unit 512 can exclude a distribution board that supplies an amount of power less than the amount of power of the supply destination to be estimated from the estimation target. The extracted identification information of the distribution board is stored in a storage area such as the RAM 413.

  The power amount change amount calculation unit 513 calculates a series of time-series changes per predetermined time from the power amount. For example, it is assumed that the measured power amount database 507 stores 1740.0 [Wh] as “2011/9/6 12:00” as the power amount of the distribution board 1. Further, the measured power amount database 507 indicates that the power amount of the distribution board 1 is 1732.0 [Wh] at “2011/9/6 12:10” and 1729. at “2011/9/6 12:20”. Assume that 3 [Wh] is recorded.

  At this time, the power amount change amount calculation unit 513 calculates the change amount per 10 minutes as the predetermined time. Specifically, the power amount change amount calculation unit 513 calculates the power amount change amount from 12:00 to 12:10 as (1732.0-1740.0) = − 8 [Wh / 10 minutes] = − 48 [ W]. Subsequently, the power amount change amount calculation unit 513 calculates the amount of change from 12:10 to 12:20 as (1729.3-1732.0) = − 2.7 [Wh / 10 minutes] = − 16 [W ] Is calculated as follows. The calculated change amount is stored in a storage area such as the RAM 413.

  The measurement sampling rate adjustment unit 514 makes the sampling rate of the supply source and the supply destination the same for the change amount calculated by the power amount change amount calculation unit 513. For example, when the amount of change in the distribution board 1 that is the supply source is about 10 minutes and the amount of change in the power sensor C1 that is the supply destination is about 5 minutes, the measurement sampling rate adjustment unit 514 The amount of power change at C1 is adjusted to the amount of power change per 10 minutes. The adjusted change amount is stored in a storage area such as the RAM 413.

  The change amount recording unit 515 records the change amount adjusted by the measurement sampling rate adjustment unit 514 in the power amount change amount table 508. For example, the change amount recording unit 515 sets the change amount at “2011/9/6 12:10” to −48 [W] and “2011/9/6 12:20” with respect to the change amount at the distribution board 1. Is recorded in the power amount change amount table 508 as -16 [W]. Further, the change amount recording unit 515 sets the change amount at “2011/9/6 12:10” to −10 [W] and “2011/9/6 12:20” for the change amount at the outlet 1. The change amount is set to −30 [W] and recorded in the power amount change amount table 508.

  The extraction unit 521 determines, from the power amount change amount table 508, the amount of change at one of the supply sources in the supply source group in which the time information of the first correspondence information and the time information of the second correspondence information are the same time. Two or more combinations with the amount of change at the supply destination at the same time are extracted. For example, for the first combination, the extraction unit 521 has a change amount −48 [W] at the distribution board 1 and a change amount −10 at the outlet 1 at the measurement time “2011/9/6 12:10”. [W] is extracted. Further, for the second combination, the extraction unit 521 uses the change amount −16 [W] on the distribution board 1 at the second measurement time “2011/9/6 12:20” and the outlet 1 The change amount −30 [W] is extracted.

  Thus, the number of data to be extracted may be two or more times for method 1 to method 3. Due to the operation of the extraction unit 521, the calculation apparatus 101 narrows down the data to be estimated, so that the time required for estimation can be shortened. The extraction result is stored in a storage area such as the RAM 413.

  The generation unit 522 generates a statistical model indicating a probability distribution that the change amount at any of the supply sources can take based on the change amount at any of the two or more supply sources extracted by the extraction unit 521. . The statistical model may be, for example, a normal distribution, a mixed normal distribution, or an empirical distribution. For example, when the statistical model is a normal distribution, the generation unit 522 generates a normal distribution using the average and standard deviation of the change amount of the distribution board 1 as parameters. In order to generate an accurate statistical model, it is preferable that the number of changes is large, but a statistical model can be generated if there are at least two changes.

  Further, the generation unit 522 determines the observed variance value and the theoretical value of variance at any of the supply sources according to the number of change amounts at any of the two or more supply sources extracted by the extraction unit 521. A statistical model showing the distribution of the probability that the ratio can take may be generated. The statistical model is, for example, a chi-square distribution. Moreover, the observed value of the variance of the change amount at any of the suppliers is a value obtained by calculating the variance directly from the actual measurement data. Further, the theoretical value of the variance of the change amount at any of the suppliers is a value derived by a mathematical formula based on the variance of the change amount at any of the suppliers and the variance of the change amount at the supplier. .

  For example, if the number of change amounts of the distribution board is 10, the generation unit 522 generates a chi-square distribution with the degree of freedom 10-1 = 9 as a parameter. With the operation of the generation unit 522, the calculation apparatus 101 can prepare a probability density function for obtaining the likelihood. The generated statistical model parameters are stored in a storage area such as the RAM 413.

  The setting unit 523 connects any one of the supply sources and the supply destination based on at least one of the change amounts at the supply source and the change amount at the supply destination. The amount of change at any of the supply sources is set in accordance with the estimated content indicating disconnection. The estimated content is either the content that is assumed that any one of the supply sources and the supply destination is connected or the content that is assumed that any one of the supply sources and the supply destination is not connected.

  Specifically, when the estimation content indicates connection, the setting unit 523 sets a value obtained by subtracting the change amount at the supply destination from the change amount at any of the supply sources, according to the estimation content indicating the connection. You may set to the variation | change_quantity in a supplier. For example, it is assumed that the change amount at the distribution board 1 at “2011/9/6 12:10” is −48 [W], and the change amount at the outlet 1 is −10 [W]. At this time, the setting unit 523 sets the amount of change of the distribution board 1 when assuming the estimation content indicating the connection as −48 − (− 10) = − 38 [W].

  Further, when the estimation content indicates disconnection, the setting unit 523 may set the change amount at any of the supply sources to the change amount at any of the supply sources according to the estimation content indicating disconnection. Good. For example, it is assumed that the change amount of the distribution board 1 at “2011/9/6 12:10” is −48 [W]. At this time, the setting unit 523 sets the amount of change of the distribution board 1 as −48 [W] when the estimated content indicating the non-connection is assumed.

  In addition, the setting unit 523 can select either one of the supply contents and the supply destination based on the estimated content indicating connection or non-connection between the supply source and the supply destination based on the change amount at the supply source and the change amount at the supply destination. The observed value of the variance of the amount of change at the supplier and the theoretical value of the variance may be set.

  Specifically, when the estimation content indicates connection, the setting unit 523 sets the variance of the value obtained by subtracting the change amount at the supply destination from the change amount at any of the supply sources in accordance with the estimation content indicating the connection. Set the observed value of the variance of the amount of change at the supplier. Subsequently, the setting unit 523 changes the value obtained by subtracting the variance of the change amount at the supply destination from the variance of the change amount at any of the supply sources, according to the estimated content indicating the connection. Set to the theoretical value of the variance of the quantity.

  For example, assume that the variance of the change amount at distribution board 1 is 1193.7, and the variance of the change amount at outlet 1 is 309.5. At this time, the setting unit 523 sets the variance of the value obtained by subtracting the amount of change at the outlet 1 from the amount of change at the distribution board 1 to the observed value of the variance of the change amount at the distribution board 1 at each time. To do. Moreover, the setting part 523 sets the theoretical value of the dispersion | distribution of the distribution board 1 at the time of assuming the presumed content which shows a connection to 1193.7-309.5 = 884.2.

  In addition, when the estimation content indicates disconnection, the setting unit 523 sets the variance of the value obtained by subtracting the change amount at the supply destination from the change amount at any of the supply sources according to the estimation content indicating disconnection. Set the observed value of the variance of the amount of change at the supplier. Subsequently, the setting unit 523 sets the value obtained by adding the variance of the change amount at the supply destination to the variance of the change amount at any of the supply sources at any of the supply sources according to the estimation content indicating the disconnection. You may set to the theoretical value of dispersion | distribution of variation | change_quantity.

  For example, assume that the variance of the change amount at distribution board 1 is 1193.7, and the variance of the change amount at outlet 1 is 309.5. At this time, the setting unit 523 calculates the variance of the value obtained by subtracting the amount of change at the outlet 1 from the amount of change at the distribution board 1 when the estimated content indicating disconnection is assumed at each time. Set to the observed value of the variance of the change in. In addition, the setting unit 523 sets the theoretical value of the distribution of the distribution board 1 when assuming the estimation content indicating non-connection to 1193.7 + 309.5 = 1503.2. By the operation of the setting unit 523, the calculation apparatus 101 can set an argument to be input to the probability model. The set information is stored in a storage area such as the RAM 413.

  The calculation unit 524 inputs the amount of change set by the setting unit 523 to the statistical model generated by the generation unit 522, and calculates a probability indicating the degree of correctness of the estimated content. Specifically, the calculation unit 524 inputs, to the statistical model generated by the generation unit 522, a change amount corresponding to the estimated content indicating the connection set by the setting unit 523, and correctness of the estimated content indicating the connection. You may calculate the probability which shows the degree of.

  For example, when the normal distribution is generated by the generation unit 522, the calculation unit 524 inputs the change amount −38 [W] of the distribution board 1 to the generated normal distribution, and corrects the estimated content indicating the connection. A probability 0.0053 indicating the degree of severity is calculated.

  In addition, the calculation unit 524 inputs, to the statistical model generated by the generation unit 522, the amount of change corresponding to the estimation content indicating disconnection set by the setting unit 523, and correctness of the estimation content indicating disconnection. You may calculate the probability which shows the degree of. For example, when the normal distribution is generated by the generation unit 522, the calculation unit 524 inputs the change amount −48 [W] of the distribution board 1 to the generated normal distribution, and the estimated content indicating disconnection. A probability of 0.0034 indicating the degree of correctness is calculated.

  In addition, the calculation unit 524 determines which one of the supply source and the supply destination is based on the probability indicating the correctness of the estimated content indicating connection and the probability indicating the correctness of the estimated content indicating non-connection. An index value indicating the possibility of connection may be calculated.

  When the supply source and the supply destination are connected, the probability indicating the degree of correctness of the estimated content indicating the connection is increased, and the probability indicating the correctness of the estimated content indicating the non-connection is decreased. Therefore, for example, the calculation unit 524 may calculate the index value as a value obtained by subtracting the probability indicating the degree of correctness of the estimated content indicating disconnection from the probability indicating the degree of correctness of the estimated content indicating connection. Good. Alternatively, the calculation unit 524 may calculate the index value as a value obtained by dividing the probability indicating the degree of correctness of the estimated content indicating connection by the probability indicating the degree of correctness of the estimated content indicating non-connection.

  For example, when the probability indicating the degree of correctness of the estimated content indicating connection is 0.0053 and the probability indicating the correctness of the estimated content indicating non-connection is 0.0034, The value is calculated as 0.0053 / 0.0034≈1.56.

  Further, the calculation unit 524 inputs the ratio of the observed dispersion value and the theoretical dispersion value set by the setting unit 523 to the statistical model generated by the generation unit 522, and determines the degree of correctness of the estimated content. The probability shown may be calculated. Specifically, the calculation unit 524 inputs the ratio of the observed value of variance and the theoretical value of variance according to the estimation content indicating the connection set by the setting unit 523 to the statistical model generated by the generating unit 522. Thus, the probability indicating the degree of correctness of the estimated content indicating the connection may be calculated.

  For example, the generation unit 522 generates a chi-square distribution with 9 degrees of freedom. Furthermore, the setting unit 523 calculates ((number of change amounts of distribution board −1) × observed value of dispersion / dispersion theory as the ratio of the observed value of dispersion and the theoretical value of dispersion in the estimation content indicating the connection. Value) = 5.19 is set. At this time, the calculation unit 524 calculates the probability indicating the degree of correctness of the estimated content indicating the connection as 0.09.

  In addition, the calculation unit 524 inputs the ratio of the observed dispersion value and the theoretical dispersion value according to the estimation content indicating the non-connection set by the setting unit 523 to the statistical model generated by the generation unit 522. The probability indicating the degree of correctness of the estimated content indicating disconnection may be calculated. For example, it is assumed that the generation unit 522 generates a chi-square distribution with 9 degrees of freedom. Further, the setting unit 523 calculates ((number of change in the distribution board minus 1) × observed value of dispersion / dispersion) as the ratio between the observed dispersion value and the theoretical dispersion value in the estimation content indicating disconnection. It is assumed that (theoretical value) = 3.05 is set. At this time, the calculation unit 524 calculates the probability indicating the degree of correctness of the estimated content indicating the connection as 0.04.

  By the operation of the calculation unit 524, the calculation apparatus 101 can obtain a probability indicating the degree of correctness of the estimated content. The calculation result is stored in a storage area such as the RAM 413.

  The calculation unit 525 selects one of the supply sources based on the sign of the change amount at any two or more supply sources extracted by the extraction unit 521 and the sign of the change amount at two or more supply destinations. And an index value indicating the possibility that the supply destination is connected.

  For example, if the change amount at the distribution board 1 at the measurement time “2011/9/6 12:10” is −48 [W] and the change amount at the outlet 1 is −10 [W], the calculation is performed. The unit 525 encodes minus which becomes the sign of the change amount in the distribution board 1 to -1. Similarly, the calculation unit 525 encodes minus, which is the sign of the amount of change at the outlet 1, into -1. Subsequently, the calculation unit 525 calculates the sign of the amount of change at the distribution board 1 and the sign of the amount of change at the outlet 1 at the measurement time “2011/9/6 12:20”. Subsequently, the calculation unit 525 may calculate the index value by summing the values obtained by multiplying the encoded values at each time. Alternatively, the calculation unit 525 may calculate the correlation coefficient of the encoded value. By the operation of the calculation unit 525, the calculation apparatus 101 can obtain the ratio at which the increase / decrease in the amount of change matches between the supply source and the supply destination. The calculation result is stored in a storage area such as the RAM 413.

  The output unit 526 outputs the probability calculated by the calculation unit 524. The output unit 526 may output the index value calculated by the calculation unit 524 or the calculation unit 525. The output destination may be output to a storage area such as the RAM 413 or the magnetic disk 415, or may be output to a device that has issued a request for estimating the correspondence between the supply source and the supply destination.

  FIG. 6 is a sequence diagram illustrating an operation example of the communication system. The power sensor B1 in the distribution board 1 transmits the measured power amount to the calculation device 101 (step S601). In addition, the power sensor C1 in the outlet 1 transmits the measured power amount to the calculation device 101 (step S602). Note that the sampling rate of the amount of power measured by the power sensor B1 in the distribution board 1 may be different from the sampling rate measured by the power sensor C1. Further, the processes in steps S601 and S602 are continuously performed.

  The measurement data receiving unit 501 receives the amount of power (step S603). Note that the process of step S603 is continuously performed. The measurement data recording unit 502 records the received power amount in the measured power amount database 507 (step S604).

  When an estimation request as to which supply source is connected to a specific supply destination is received from the user, the change amount generation unit 503 reads the power amount from the measured power amount database 507 (step S605). Next, the change amount generation unit 503 calculates a power amount change amount (step S606). Subsequently, the change amount generation unit 503 adjusts the measurement sampling rate (step S607). After the execution of step S607 is completed, the change amount generation unit 503 records the change amount in the power amount change amount table 508 (step S608).

  When the constant change amount is accumulated in the power amount change amount table 508, the connection relationship estimation unit 504 reads the change amount from the power amount change amount table 508 (step S609). Next, the connection relationship estimation unit 504 estimates the connection relationship (step S610). Subsequently, the connection relationship estimation unit 504 outputs an index value that is an estimation result (step S611). After the index value is output by the connection relationship estimation unit 504, the index value recording unit 505 records the index value (step S612).

  FIG. 7 is an explanatory diagram showing an example of the contents stored in the measured power amount database. The measured power amount database 507 includes a supply source power amount table 701 and a supply destination power amount table 702. The supply power amount table 701 illustrated in FIG. 7 stores records 701-1 to 701-3. In addition, the supply power amount table 702 illustrated in FIG. 7 stores records 702-1 to 702-3.

  The power supply amount table 701 includes fields of time and power amount at each supply source. In the time field, the time when the electric energy is measured is stored. The amount of power for each supply source is stored in the power amount field for each supply source. The supply destination power amount table 702 includes fields of time and the amount of power at each supply destination. In the time field, the time when the electric energy is measured is stored. The amount of power for each supply destination is stored in the power amount field for each supply destination.

  For example, in the record 701-1, the electric energy in the distribution board 1 in “2011/9/6 12:00” is 1740.0 [Wh], and the electric energy in the distribution board 2 is 4500.0 [Wh]. ]. In addition, in the record 702-1, the power amount at the outlet 1 in “2011/9/6 12:00” is 88.0 [Wh], and the power amount at the outlet 2 is 48.0 [Wh]. It shows that the electric energy of the outlet 3 is 0.7 [Wh].

  FIG. 8 is an explanatory diagram of an example of the contents stored in the power amount change amount table. The power amount change amount table 508 includes a supply source power amount change amount table 801 and a supply destination power amount change amount table 802. The supply source power amount change amount table 801 illustrated in FIG. 8 stores a record 801-1 and a record 801-2. Further, the supply power amount change amount table 802 illustrated in FIG. 8 stores records 802-1 and records 802-2. The first correspondence information is each record of the supply source power amount change amount table 801. Specifically, in FIG. 8, the first correspondence information is a record 801-1 and a record 801-2. The second correspondence information is each record of the supply destination power amount change amount table 802. Specifically, in FIG. 8, the second correspondence information is a record 802-1 and a record 802-2.

  The supply source power amount change amount table 801 includes fields of time and change amount at each supply source. The time field stores the time after the change in the measurement time of the change amount. The change amount field at each supplier stores the change amount until a predetermined time elapses from the previous time at each supplier. In the present embodiment, the predetermined time is 10 minutes. Further, the supply power amount change amount table 802 includes fields of time and change amount at the supply destination. The time field stores the time after the change in the measurement time of the change amount. The change amount field at the supply destination stores a change amount until a predetermined time elapses from the previous time at the supply destination.

  For example, in the record 801-1, the change in power from 10 minutes before “2011/9/6 12:10” is −48 [W] for the distribution board 1 and 20 [W] for the distribution board 2. W]. The record 802-1 indicates that the amount of change in power from 10 minutes before “2011/9/6 12:10” is −10 [W] at the outlet 1.

  Note that the change amount generation unit 503 generates the power amount change amount table 508 with reference to the measured power amount database 507. For example, as shown below, the change amount generation unit 503 uses the value 1732.0 [Wh] of the power amount field in the distribution board 1 of the record 701-2 to calculate the change amount in the distribution board 1 of the record 701-1. The value 1740.0 [Wh] in the electric energy field is subtracted.

  Change in electric energy at 12:00 to 12:10 of distribution board 1 = 1732.0 [Wh] -1740.0 [Wh] = − 8 [Wh / 10 minutes] = − 48 [W]

  The change amount generation unit 503 stores the calculated −48 [W] in the power amount change amount field in the distribution board 1 of the record 801-1. The change amount generation unit 503 similarly calculates the power amount change amount in the distribution board 2 and the outlet 1 and stores the same in the power amount change amount table 508.

Next, a method by which the connection relationship estimation unit 504 estimates the connection relationship will be described with reference to the power amount change amount table 508 shown in FIG. In the present embodiment, Method 1 to Method 3 for estimating the connection relationship will be described with reference to FIGS. 9A to 13. In FIG. 9A to FIG. 13, an example is described in which a specific supply destination is the outlet 1 and an index value indicating the possibility of connection with respect to the distribution board 1 in the supply source group is output. Also, let C (t _k ) be the amount of power change at outlet 1 at time t _k . Further, let B (t _k ) be the amount of change in electric power at the distribution board 1. k is an integer from 1 to 10. Further, it is assumed that the power amount change amount table 508 stores the power amount change amounts of the outlet 1 and the distribution board 1 from time t ₁ to time t ₁₀ .

  9A and 9B are explanatory diagrams illustrating an operation example of the method 1 in the connection relationship estimation unit. A graph 901 in FIG. 9A shows the amount of change in electric energy between the distribution board 1 and the outlet 1. The horizontal axis of the graph 901 indicates time, and the vertical axis indicates the amount of power change. A table 902 shows an example of an extraction result from the power amount change amount table 508. Specifically, the table 902 shows the amount of power change at the outlet 1 and the amount of power change at the distribution board 1 at each time. Table 903 shows an example of the conversion result from the extraction result to the sign of the amount of power change. Specifically, the table 903 shows the sign of the amount of power change at the outlet 1 at each time and the sign of the amount of power change at the distribution board 1. Further, the table 904 shows an example of the calculated index value.

As shown in Table 902, the connection relationship estimation unit 504 extracts the power amount change amount at the time when the absolute value of the power amount change amount is equal to or greater than the threshold CV for the power amount change amount of the distribution board 1 and the outlet 1. . In addition, the connection relationship estimation unit 504 extracts the time when the amount of power change of the distribution board 1 and the outlet 1 existed at the same time. As shown in the graph 901, the amount of power variation in the outlet 1 at time t ₈ becomes less than the threshold value CV. Further, the amount of power variation at the time t ₁₀ in the distribution panel 1 becomes less than the threshold value CV. Therefore, the connection relationship estimation unit 504 extracts the amount of power change between the distribution board 1 and the outlet 1 from time t ₁ to time t ₇ and time t ₉ .

Next, as shown in Table 903, the connection relationship estimation unit 504 indicates that the distribution board 1 and the outlet 1 may be connected from the sign of the amount of power change between the distribution board 1 and the outlet 1. The index value shown is calculated. Here, the sign of the outlet 1 at each time is S _C (t _k ), and the sign of the distribution board 1 is S _B (t _k ). For example, the sign of the outlet 1 can be calculated by S _C (t _k ) = Sign (C (t _k )). Here, Sign () is a function that returns 1 if the argument is positive and returns -1 if the argument is negative.

Subsequently, as shown in Table 904, the connection relationship estimation unit 504 calculates the correlation coefficient R ₁ of S _C (t _k ) and S _B (t _k ) as index values. For example, the connection relationship estimation unit 504 calculates a correlation coefficient R ₁ = 0.7745. In the method 1, the correlation coefficient R is an index value indicating the possibility that the distribution board and the outlet are connected. Further, the connection relationship estimation unit 504 calculates a correlation coefficient R ₂ between the distribution board 2 and the outlet 1. Subsequently, the result output unit 506 outputs the identification information of the distribution board having the largest correlation coefficient, for example.

  As shown in FIGS. 9A and 9B, the connection relationship estimation unit 504 increases and decreases the power amount of the outlet because the power amount of the connected distribution board often increases as the power amount of the outlet increases. The correlation coefficient R is calculated by quantifying the rate at which the increase / decrease in the power amount of the distribution board matches.

  10A and 10B are explanatory diagrams illustrating an operation example of the method 2 in the connection relationship estimation unit. Method 2 is a method of calculating a likelihood when a connection is assumed and a likelihood when a non-connection is assumed using a chi-square distribution to obtain a connection relationship having a high likelihood. Also, in method 2, assuming that the amount of change at each time is independent, the likelihood is calculated by substituting the ratio of the observed value of variance of the amount of change in power and the theoretical value of variance into the chi-square distribution. To do.

  A graph 1001 shows the amount of change in electric power between the distribution board 1 and the outlet 1. The horizontal axis of the graph 1001 indicates time, and the vertical axis indicates the amount of change in electric energy. A table 1002 shows an example of an extraction result from the power amount change amount table 508. Specifically, a table 1002 shows the amount of power change at the outlet 1 and the amount of power change at the distribution board 1 at each time. A graph 1003 is a graph of the probability density function of the chi-square distribution. The horizontal axis of the graph 1003 is the value of the ratio of the observed value of variance in the amount of change in electric energy and the theoretical value of variance. The vertical axis of the graph 1003 indicates the probability that the ratio value can take. A table 1004 shows an example of the calculated index value.

As shown in the graph 1001 and the table 1002, the connection relationship estimation unit 504 extracts the time when the amount of power change between the distribution board 1 and the outlet 1 exists at the same time. In the example of FIG. 10, there is a power amount change amount at all times from time t ₁ to time t ₁₀ . Therefore, the connection relationship estimation unit 504 extracts the amount of change in the amount of power between the distribution board 1 and the outlet 1 from time t ₁ to time t ₁₀ . The state where there is no power amount change is a case where the power sensor cannot measure the time for some reason. The cause of the failure to measure is, for example, when the firmware of the power sensor C1 in the outlet 1 is being updated and the power sensor C1 cannot measure the amount of power.

After extracting the power amount change amount, the connection relationship estimation unit 504 calculates σ _C ² as the variance of the power amount change amount per predetermined time at the outlet 1. Similarly, the connection relationship estimation unit 504 calculates σ _B ² as the variance of the amount of power change per predetermined time on the distribution board 1. The calculation method of σ _C ² and σ _B ² may be sample variance or unbiased variance. For example, when calculating the variance at the outlet 1, the connection relationship estimation unit 504 calculates the variance by the following equation (1) if the sample variance is adopted, and by the following equation (2) if the unbiased variance is adopted. Calculate the variance.

Here, μ _C is the average of the amount of power change per predetermined time at the outlet 1. Further, n is the number of extracted power amount changes at each time. For example, when extraction is performed for times t _{1 to} t ₁₀ , n = 10. Hereinafter, it is assumed that the connection relationship estimation unit 504 performs the variance calculation method using equation (2).

Next, the connection relationship estimation unit 504 calculates the theoretical value σ _BC ² of the variance of the amount of power change of the distribution board 1 excluding the amount of power change at the outlet 1 when connection is assumed. Specifically, the connection relationship estimation unit 504 calculates σ _BC ² using the following equation (3).

σ _BC ² = σ _B ² −σ _C ² (3)

Since connection is assumed in the expression (3), σ _BC ² is a value obtained by subtracting σ _C ² from σ _B ² . For example, assuming that σ _B ² = 1193.7 and σ _C ² = 309.5, the connection relationship estimation unit 504 calculates σ _BC ² using Equation (3) as follows.

σ _BC ² = 1193.7-309.5 = 884.2

Subsequently, the connection relationship estimation unit 504 obtains the theoretical value σ _BN ² of the variance of the power amount change amount of the distribution board 1 excluding the power amount change amount at the outlet 1 when non-connection is assumed (4) Calculate using the formula.

σ _BN ² = σ _B ² + σ _C ² (4)

(4) Since it is assumed that connection is not established, the amount of change in distribution board 1 and the amount of change in outlet 1 are independent, and σ _BN ² is a value obtained by adding σ _C ² to σ _B ² Become. For example, assuming that σ _B ² = 1193.7 and σ _C ² = 309.5, the connection relationship estimation unit 504 calculates σ _BN ² using Equation (4) as follows.

σ _BN ² = 1193.7 + 309.5 = 1503.2

Next, the connection relationship estimation unit 504 calculates the observed value σ _real ² of the variance of the power amount change amount of the distribution board 1 excluding the power amount change amount at the outlet 1. σ _real ² is a variance of values obtained by subtracting the power amount change amount of the outlet 1 from the power amount change amount of the distribution board 1 at each time.

Subsequently, the connection relationship estimation unit 504 sets the degree of freedom of the chi-square distribution. The degree of freedom is a number obtained by subtracting 1 from the number of extracted power amount changes at each time. In the case of FIG. 10, 10-1 = 9. Next, the connection relationship estimation unit 504 uses the following equation (5) to calculate the ratio Z _BC between the observed value of variance in the amount of change in the electric energy of the distribution board 1 and the theoretical value of variance when connection is assumed. Calculate as follows.

Z _BC = (n−1) × σ _real ² / σ _BC ² (5)

For example, assuming that n = 10, σ _BC ² = 884.2, and σ _real ² = 573.7, the connection relationship estimation unit 504 calculates Z _BC as follows using equation (5).

Z _BC = (10-1) × 573.7 / 884.2 = 5.19

In addition, the connection relationship estimation unit 504 uses the following equation (6) to calculate the ratio Z _BN between the observed dispersion value and the theoretical dispersion value of the distribution amount of the distribution board 1 assuming no connection. Calculate as follows.

Z _BN = (n−1) × σ _real ² / σ _BN ² (6)

For example, assuming that n = 10, σ _BN ² = 1503.2, and σ _real ² = 573.7, the connection relationship estimation unit 504 calculates Z _BN using Equation (6) as follows.

Z _BN = (10-1) × 573.7 / 1503.2 = 3.05

Next, the connection relationship estimation unit 504 calculates the probability P _C (σ _real ² ) that σ _real ² is observed when connection is assumed, using the following equation (7).

P _C (σ _real ² ) = f (Z _BC ; n−1) (7)

Here, f is a probability density function of a chi-square distribution. Here, a graph 1003 shows a probability density function of a chi-square distribution with 9 degrees of freedom. For example, if Z _BC = 5.19 and n = 10, then P _C (σ _real ² ) = 0.09.

Further, the connection relationship estimation unit 504 calculates the probability P _N (σ _real ² ) that σ _real ² is observed when non-connection is assumed, using the following equation (8).

P _N (σ _real ² ) = f (Z _BN ; n−1) (8)

For example, if Z _BN = 3.05 and n = 10, then P _N (σ _real ² ) = 0.04. Subsequently, after calculating the probabilities using the equations (7) and (8), the connection relationship estimation unit 504 sets an index value R ₁ indicating the possibility that the distribution board 1 and the outlet 1 are connected to the following (9 ) To calculate.

R ₁ = P _C (σ _real ² ) / P _N (σ _real ² ) (9)

The calculation result using the formula (9) is the value shown in Table 1004, for example. Specifically, the connection relationship estimation unit 504 calculates R ₁ = 0.09 / 0.04 = 2.20 as an index value.

  As described above, in FIGS. 10A and 10B, the connection relationship between the distribution board and the outlet can be estimated without following the wiring by obtaining the likelihood when connection or disconnection is assumed.

  11A and 11B are explanatory diagrams (part 1) illustrating an operation example of the method 3 in the connection relationship estimation unit. Method 3 is a method of calculating the likelihood when connection is assumed and the likelihood when connection is assumed, using the probability density function of the distribution board. 11A and 11B illustrate an example in which the amount of change in electric power that can be generated in the distribution board follows a normal distribution. FIG. 12 illustrates an example in which the amount of change in electric power that can be generated in the distribution board follows a mixed normal distribution. FIG. 13 illustrates an example in which the amount of change in electric power that can occur in the distribution board follows an empirical distribution.

  A graph 1001 shows the amount of change in electric power between the distribution board 1 and the outlet 1. A table 1002 shows an example of an extraction result from the power amount change amount table 508. A graph 1101 is a graph of a probability density function that is a probability distribution of the amount of change in the amount of power in the distribution board 1. Table 1102 shows the probability that the amount of power change when connection is assumed and the probability of change of power amount when connection is assumed at each time. Table 1103 shows an example of the calculated index value.

As shown in the graph 1001 and the table 1002, the connection relationship estimation unit 504 extracts the time when the amount of power change between the distribution board 1 and the outlet 1 exists at the same time. Since this extraction method is the same as the method shown in FIG. 10A, description thereof is omitted. Subsequently, the connection relationship estimation unit 504 calculates the average μ _B of the amount of change in electric power per predetermined time in the distribution board 1. Further, the connection relationship estimation unit 504 calculates a standard deviation σ _B of the amount of change in electric power per predetermined time in the distribution board 1.

  Next, the connection relationship estimation unit 504 generates a probability density function of the amount of power change of the distribution board 1. A graph 1101 is a graph of a probability density function when it is assumed that the amount of power change that can be generated in the distribution board 1 follows a normal distribution. The horizontal axis of the graph 1101 indicates the amount of power change. In addition, the vertical axis of the graph 1101 indicates the occurrence probability of the power amount change amount. The connection relationship estimation unit 504 calculates the likelihood LC when the connection is assumed and the likelihood LN when the connection is assumed, using the probability distribution of the amount of change in the electric power of the distribution board 1.

The probability LC = P _C (B ₁ = B ₁ (t _k ) | C = C (t _k )) that the amount of change in the electric energy of the distribution board 1 when connection is assumed is expressed by the following equation (10): It can be approximated as follows.

LC = P _C (B ₁ = B ₁ (t _k ) | C = C (t _k )) ≈f (B (t _k ) −C (t _k )) = f _C (B (t _k )) ( 10)

  Here, f is a probability density function indicating a probability distribution that can be taken by the amount of power change in the distribution board 1. As a reason that can be approximated as in equation (10), when connection is assumed, a distribution destination other than the outlet 1 is also connected to the distribution board 1, but the amount of power change at the supply destination other than the outlet 1 Is independent of the amount of power change of the outlet 1. Therefore, it can be assumed that the probability density function of the power amount change amount of the distribution board 1 excluding the power amount change amount at the outlet 1 substantially matches the probability density function of the distribution board 1. Such approximation can be performed.

For example, the connection relationship estimation unit 504 obtains the probability LC that the amount of power change of the distribution board 1 at time t ₁ when the connection is assumed, using the equation (10) as follows.

LC = P _C (B ₁ = B ₁ (t ₁ ) | C = C (t ₁ )) ≈f (B (t ₁ ) −C (t ₁ )) = f (−48 − (− 10)) ≈ 0.0053

Further, the probability that the amount of power change of distribution board 1 is observed excluding the amount of power change at outlet 1 assuming no connection LN = P _N (B ₁ = B ₁ (t _k ) | C = C (t _k )) is expressed by the following equation (11).

LN = P _N (B ₁ = B ₁ (t _k ) | C = C (t _k )) = f (B (t _k )) = f _N (B (t _k )) (11)

Since it is assumed that connection (11) is not connected, the probability that the amount of power change of distribution board 1 is observed is f (B (t _k )) regardless of the amount of power change of outlet 1. It becomes. For example, the connection relationship estimation unit 504 obtains the probability LN that the amount of change in the power amount of the distribution board 1 at time t ₁ when no connection is assumed, using the equation (11) as follows.

LN = P _N (B ₁ = B ₁ (t ₁ ) | C = C (t ₁ )) ≈f (B (t ₁ ) −C (t ₁ )) = f (−48) ≈0.0034

Table 1102 is an example of calculating the probability that the amount of change in electric power from time t ₁ to time t ₁₀ is observed. Here, the reason why the likelihood when assuming the connection when the outlet 1 and the distribution board 1 are actually connected will be described.

When outlet 1 and distribution board 1 are connected and C (t _k ) is a positive value, f _c (t _k ) is a positive direction from f _N (t _k ) as shown by graph 1101. In many cases, B (t _k ) also takes a positive value. Therefore, LC (t _k ) tends to be larger than LN (t _k ). Further, when the outlet 1 and the distribution board 1 are connected and C (t _k ) is a negative value, the function in which f _c (t _k ) moves from f _N (t _k ) in the negative direction Therefore, B (t _k ) often takes a negative value. Therefore, LC (t _k ) tends to be larger than LN (t _k ).

On the other hand, when outlet 1 and distribution board 1 are not connected and C (t _k ) is a positive value, f _c (t _k ) is a function that has moved in the positive direction from f _N (t _k ). However, B (t _k ) is a value independent of C (t _k ), and the ease of taking a positive value and a negative value is the same. Therefore, LC (t _k ) is not necessarily larger than LN (t _k ). Similarly, when the outlet 1 and the distribution board 1 are not connected and C (t _k ) is a negative value, LC (t _k ) is not always greater than LN (t _k ). Therefore, when the outlet 1 and the distribution board 1 are actually connected, the likelihood when the connection is assumed increases.

After calculating the probabilities using the equations (10) and (11), the connection relationship estimation unit 504 uses the following equation (12) as an index value R ₁ indicating the possibility that the distribution board 1 and the outlet 1 are connected. Use to calculate.

For example, as shown in Table 1103, the connection relationship estimation unit 504 calculates R ₁ using Equation (12) as follows.

R ₁ = (LC (t ₁ ) / LN (t ₁ )) × (LC (t ₂ ) / LN (t ₂ )) ××× (LC (t ₁₀ ) / LN (t ₁₀ )) = 7.9403

  As described above, the operation of the connection relationship estimation unit 504 illustrated in FIG. 11 allows the calculation apparatus 101 to obtain the likelihood when connection or disconnection is assumed, so that the distribution board and the outlet can be connected without following the wiring. Connection relations can be estimated.

  FIG. 12 is an explanatory diagram (part 2) of an operation example of the method 3 in the connection relationship estimation unit. FIG. 12 illustrates an example in which the amount of change in electric power that can be generated in the distribution board follows a mixed normal distribution. Here, the mixed normal distribution is a probability distribution expressed by a weighted sum of normal distributions. The total weight is 1. A graph 1201 shows an example of a mixed normal distribution. The horizontal axis of the graph 1201 indicates the amount of power change. The vertical axis of the graph 1201 indicates the occurrence probability of the power amount change amount.

  The probability density function f (x) of the mixed normal distribution is expressed by the following equation (13), where m is the number of normal distributions.

Here, μ _j represents the average of the j-th normal distribution. Σ _j ² represents the variance of the j-th normal distribution. Ξ _j represents the weight of the j-th normal distribution. Φ (x; μ _j , σ _j ² ) represents a probability density function of the jth normal distribution. The number of normal distributions may be any number, but in this embodiment, for example, m = 3 is sufficient.

Parameters such as weight ξ _j , average μ _j , and variance σ _j ^{2 of} each normal distribution can be obtained by, for example, an EM (Expectation Maximization) algorithm. A flowchart for calculating parameters using the EM algorithm will be described later with reference to FIG.

  The reason why the mixed normal distribution is used is that there is a case where the amount of power change in the distribution board does not follow the normal distribution. For example, this is a case where the amount of power change in the distribution board has a plurality of peaks. In such a case, since the connection relationship estimation unit 504 can generate a probability density function that is closer to the actually measured value, the likelihood calculated using the mixed normal distribution is the likelihood calculated using the normal distribution. Will be more accurate.

  FIG. 13 is an explanatory diagram (part 3) of an operation example of the method 3 in the connection relationship estimation unit. An example in which the amount of power change that can be generated in the distribution board follows an empirical distribution will be described. Here, the experience distribution is a probability distribution in which the area of the histogram of the amount of change in electric power is 1. A graph 1301 shows an example of an experience distribution. The horizontal axis of the graph 1301 indicates the amount of power change. In addition, the vertical axis of the graph 1301 indicates the probability of occurrence of the power amount change amount.

  Specifically, the connection relationship estimation unit 504 calculates a probability density function in the class j using the following equation (14).

However, C _j denotes the frequency of in class j. W indicates the width of the class. Q represents the number of classes. The reason for using the empirical distribution is that there is a case where the amount of change in the electric energy at the distribution board does not follow the normal distribution. In such a case, since the connection relationship estimation unit 504 can generate a probability density function closer to the actual measurement value, the likelihood calculated using the empirical distribution is higher than the likelihood calculated using the normal distribution. Is also an accurate value.

  14A and 14B are explanatory diagrams illustrating an output example in the result output unit. 14A and 14B show (A) in FIG. 14A to (D) in FIG. 14B as output examples of the result output unit 506. The output example shown in (A) of FIG. 14A shows information that “the distribution board that may be connected to the outlet 1 is the distribution board 1”. The output example shown in (A) of FIG. 14A is output when, for example, an index value that is equal to or greater than a specific threshold among the index value group output by the connection relationship estimation unit 504 is the distribution board 1. .

  In the output example shown in FIG. 14A (B), “the distribution boards that may be connected to the outlet 1 are the distribution board 1 and the distribution board 2 in descending order of possibility”. Indicates information. The output example shown in (B) of FIG. 14A is, for example, the distribution panel 1 and the distribution board 2 that have a specific threshold value or more among the index value group output by the connection relationship estimation unit 504. The index value is output when the distribution panel 1 is larger.

  The output example shown in (C) of FIG. 14A shows information that “the distribution board that may be connected to the outlet 1 could not be estimated”. The output example illustrated in (C) of FIG. 14A is output, for example, when there is no index value that is equal to or greater than a specific threshold among the index value group output by the connection relationship estimation unit 504.

The output example shown in (D) of FIG. 14B is a diagram illustrating an estimation result of the connection relation between distribution board 1 and distribution board 2 that are supply sources in power sensor network 100 and outlet 1 to outlet 3 that are supply destinations. It is an example of showing and outputting. In the output example shown in (D) of FIG. 14B, the result output unit 506 outputs the connection relationship in which the index value R is 2 or more as a specific threshold value with a solid line. In addition, the result output unit 506 connects connection relationships in which the index value R is less than 2 with broken lines. For example, since the index value R ₁ _C1 indicating the possibility that the distribution board 1 and the outlet 1 are connected is 2.20, which is 2 or more, the result output unit 506 connects the distribution board 1 and the outlet 1 to each other. Output in the state of being connected with a solid line. Moreover, since the index value R ₂ _C1 indicating the possibility that the distribution board 2 and the outlet 1 are connected is 0.41 and less than 2, the result output unit 506 connects the distribution board 2 and the outlet 1 to each other. The output is shown in a state connected by a broken line.

  Note that all estimation results may be less than a specific threshold. In this case, the result output unit 506 may output an output example as shown in (C) of FIG. 14A. Alternatively, the result output unit 506 may display that the user is inquired whether or not to perform re-estimation by changing the number of data used for measurement. Specifically, when the connection relationship estimation unit 504 estimates the amount of power change every 10 minutes from the data for one day, the result output unit 506 reads the data every 10 minutes from the data for two days or more. Outputs an inquiry as to whether or not to estimate using the amount of change in electric energy.

  Next, a flowchart of the connection relationship estimation process executed by the connection relationship estimation unit 504 illustrated in FIGS. 9A to 13 will be described.

  FIG. 15 is a flowchart illustrating an example of the method 1 in the connection relationship estimation processing procedure. FIG. 15 shows a flowchart of the method 1 shown in FIG. The calculation apparatus 101 sets the variable i to 1 (step S1501). Next, the calculation apparatus 101 selects the distribution board i as an estimation object among the distribution boards 1 to N (step S1502). Subsequently, the calculation apparatus 101 extracts, from the power amount change amount table 508, the amount of change at the time when both the distribution panel i and the outlet have values whose absolute values are equal to or greater than the threshold value CV (step S1503).

Next, the calculation apparatus 101 encodes the extracted amount of change at the distribution board i and outlet at each time (step S1504). Subsequently, the calculation apparatus 101 calculates a correlation coefficient R _i related to the distribution board i and the outlet for the encoded change amount at each time (step S1505). Next, the calculation apparatus 101 outputs a correlation coefficient R _i (step S1506).

  Subsequently, the calculation apparatus 101 determines whether or not estimation has been performed for all distribution boards (step S1507). When the estimation is not performed for all distribution boards (step S1507: No), the calculation apparatus 101 increments the variable i (step S1508). After executing the process of step S1508, the calculation apparatus 101 proceeds to the process of step S1502. When estimating with respect to all the distribution boards (step S1507: Yes), the calculation apparatus 101 complete | finishes a connection relationship estimation process. In this way, the connection relationship estimation process shown in FIG. 15 can output the index value and provide the connection source estimation result for the connection destination to the user. Further, the connection relationship estimation processing according to the method 1 has an effect of outputting an index value that is hardly influenced by an exceptional change amount even if an exceptional change amount that does not normally occur is generated.

  16A and 16B are flowcharts illustrating an example of method 2 in the connection relationship estimation processing procedure. 16A and 16B show a flowchart of the method 2 shown in FIGS. 10A and 10B. Further, step S1601 and step S1602 shown in FIG. 16A are the same as step S1501 and step S1502, and thus the description thereof is omitted.

In FIG. 16A, after executing the process of step S1602, the calculation apparatus 101 extracts, from the power amount change amount table 508, the change amount of time at which both values exist for the distribution board i and the outlet (step S1603). Subsequently, the calculation apparatus 101 sets the degree of freedom to the number of change amounts of the extracted distribution board i minus 1 (step S1604). Next, the calculation apparatus 101 calculates the variance σ _C ² of the change amount per predetermined time of the outlet (step S1605). Subsequently, the calculation apparatus 101 calculates the variance σ _B ² of the change amount per predetermined time of the distribution board i (step S1606). The calculation formula in step S1605 and step S1606 is, for example, formula (2).

Next, the calculating apparatus 101 calculates the theoretical value σ _BC ² of the variance of the power amount change amount of the distribution board i excluding the power amount change amount at the outlet when the connection is assumed (step S1607). The calculation formula in step S1607 is formula (3). Subsequently, the calculating apparatus 101 calculates the theoretical value σ _BN ² of the variance of the power amount change amount of the distribution board i excluding the power amount change amount at the outlet when the connection is assumed (step S1608). The calculation formula in step S1608 is formula (4). Next, the calculation apparatus 101 calculates the observed value σ _real ² of the variance of the power amount change amount of the distribution board i excluding the power amount change amount at the outlet (step S1609).

Next, in FIG. 16B, the calculation apparatus 101 calculates a ratio Z _BC between the observed dispersion value and the theoretical dispersion value of the distribution amount i of the distribution board i when connection is assumed (step S1610). The calculation formula in step S1610 is formula (5). Subsequently, the calculation apparatus 101 calculates the ratio Z _BN of the observed dispersion value and the theoretical dispersion value of the distribution amount i of the distribution board i assuming no connection (step S1611). The calculation formula in step S1611 is formula (6). Next, the calculation apparatus 101 generates a χ square distribution f (x; n−1) having n−1 degrees of freedom (step S1612).

Subsequently, using the generated χ-square distribution, the calculation apparatus 101 calculates a probability indicating the degree of correctness of the estimated content indicating the connection by using the χ-square distribution f (Z _BC ; n−1) (step S1613). ). The calculation formula in step S1613 is formula (7). Next, the calculation apparatus 101 calculates a probability indicating the degree of correctness of the estimated content indicating non-connection based on the χ-square distribution f (Z _BN ; n−1) (step S1614). The calculation formula in step S1613 is formula (8).

Subsequently, the calculation apparatus 101 calculates the index value R _i of the distribution board i (step S1615). The calculation formula in step S1615 is formula (9). Next, the calculation apparatus 101 outputs the index value R _i (step S1616).

  Subsequently, the calculation apparatus 101 determines whether or not estimation has been performed for all distribution boards (step S1617). When the estimation has not been made for all distribution boards (step S1617: No), the calculation apparatus 101 increments the variable i (step S1618). After completing the execution of step S1618, the calculation device 101 proceeds to the process of step S1602. When estimating with respect to all the distribution boards (step S1617: Yes), the calculation apparatus 101 complete | finishes a connection relationship estimation process.

  In this way, the connection relationship estimation processing shown in FIGS. 16A and 16B can output an index value and notify the user of the estimation result of the connection source for the connection destination. In addition, since the connection relationship estimation processing by the method 2 has a small amount of calculation, the processing result can be notified to the user quickly. Further, the connection relationship estimation processing according to the method 2 can obtain a more accurate estimation result when the variation in variation is large.

  17A and 17B are flowcharts illustrating an example of method 3 in the connection relationship estimation processing procedure. 17A and 17B show a flowchart of the method 3 shown in FIGS. 11A and 11B. Also, Steps S1701 to S1703 shown in FIG. 17A are the same as Steps S1601 to S1603, and thus description thereof is omitted.

After the execution of step S1703, the calculation apparatus 101 sets the likelihood LCi to 1 and the likelihood LNi to 1 (step S1704). Next, the calculation apparatus 101 calculates the average μ _B of the amount of change in electric power per predetermined time on the distribution board i (step S1705). Subsequently, the calculation apparatus 101 calculates the standard deviation σ _B predicted based on the sample of the amount of change in electric power per predetermined time on the distribution board i (step S1706). Next, the calculation apparatus 101 generates a probability density function f indicating a probability distribution that can be taken by the amount of change in the electric energy at the distribution board i (step S1707).

In step S1707, for example, when it is assumed that the amount of change in electric energy that can occur in the distribution board i follows a normal distribution, the normal distribution is calculated using the average μ _B and the standard deviation σ _B calculated in steps S1705 and S1706. Probability density function f according to the above is generated. When it is assumed that the amount of power change that can be generated in the distribution board i follows a mixed normal distribution, the calculation device 101 sets parameters of average value, variance, and weight as described with reference to FIG. A density function f is generated. When it is assumed that the amount of power change that can be generated in the distribution board i follows the empirical distribution, the calculation device 101 generates the probability density function f described with reference to FIG.

Next, in FIG. 17B, the calculation apparatus 101 sets 1 to k (step S1708). Subsequently, the calculating apparatus 101 calculates the likelihood LC (t _k ) at time t _k (step S1709). The calculation formula in step S1709 is formula (10). Next, the calculating apparatus 101 calculates the likelihood LN (t _k ) at time t _k (step S1710). The calculation formula in step S1710 is formula (11).

Subsequently, the calculating apparatus 101 sets LCi × LC (t _k ) for LCi (step S1711). Further, the calculation apparatus 101 sets LNi × LN (t _k ) for LNi (step S1712). Subsequently, the calculation apparatus 101 determines whether k is n or less (step S1713). When k is n or less (step S1713: Yes), the calculation apparatus 101 increments k (step S1714). After executing the process of step S1714, the calculation apparatus 101 proceeds to the process of step S1709. When k is larger than n (step S1713: No), the calculation apparatus 101 calculates the index value R _i of the distribution board _i as LCi / LNi (step S1715). Next, the calculation apparatus 101 outputs the index value R _i (step S1716).

  Subsequently, the calculation apparatus 101 determines whether or not estimation has been performed for all distribution boards (step S1717). When there is a distribution board that has not been estimated yet (step S1717: No), the calculation apparatus 101 increments the variable i (step S1718). After completing the execution of step S1718, the calculation apparatus 101 proceeds to the process of step S1702. When estimating with respect to all the distribution boards (step S1717: Yes), the calculation apparatus 101 complete | finishes a connection relationship estimation process.

  As described above, the connection relationship estimation processing shown in FIGS. 17A and 17B can output an index value and notify the user of the estimation result of the connection source for the connection destination. Further, the connection relationship estimation processing by the method 3 can obtain a more accurate result than the method 1 and the method 2.

FIG. 18 is a flowchart illustrating an example of a mixed normal distribution parameter setting processing procedure. In FIG. 18, when it is assumed that the amount of power change that can occur in the distribution board in Method 3 follows a mixed normal distribution, a flowchart for setting parameters such as weight ξ _j , average μ _j , and variance σ _j ² according to the EM algorithm. Indicates. In the following, σ _j ² is expressed as σ _j ^ 2. ^ Is a symbol indicating a power.

First, the calculation apparatus 101 sets a variable t indicating the number of parameter repetitions to 0 (step S1801). Next, the calculation apparatus 101 sets μ _j ⁽⁰⁾ , σ _j ^ 2 ⁽⁰⁾ , ξ _j ⁽⁰⁾ as parameters at t = 0 (step S1802). Note that μ _j ⁽⁰⁾ , σ _j ^ 2 ⁽⁰⁾ , and ξ _j ⁽⁰⁾ may be set to any value. J is an index of a mixed normal distribution. For example, if the number m of normal distributions is 3, j is an integer of 1 to 3. If m = 3, the calculation apparatus 101 sets a total of nine initial parameters for three initial values of μ _j ⁽⁰⁾ , σ _j ^ 2 ⁽⁰⁾ , and ξ _j ⁽⁰⁾ .

Subsequently, the computing device 101, for each Bt _k, Bt _k calculates the conditional probability Wjt _k ^(t) belonging to the mixed normal distribution j (step S1803). Wjt _k ^(t) is calculated as follows using the following equation (15).

Subsequently, the calculation apparatus 101 updates ξ _j ^{(t + 1)} (step S1804). Specifically, the calculation device 101 updates ξ _j ^{(t + 1)} using the following equation (16) as follows.

Next, the calculation apparatus 101 updates μ _j ^{(t + 1)} (step S1805). Specifically, the calculation apparatus 101 updates μ _j ^{(t + 1)} using the following equation (17) as follows.

Subsequently, the calculation apparatus 101 updates σ _j ^ 2 ^{(t + 1)} (step S1806). Specifically, the calculation apparatus 101 updates σ _j ^ 2 ^{(t + 1)} as follows using the following equation (18).

  Next, the calculation apparatus 101 determines whether or not the difference between the parameter at t and the parameter at t + 1 is equal to or smaller than a threshold value (step S1807). When larger than the threshold value (step S1807: No), the calculation apparatus 101 increments t (step S1808). After completing the execution of step S1808, the calculation apparatus 101 proceeds to the process of step S1803. If the value is equal to or smaller than the threshold value (step S1807: YES), the calculation apparatus 101 outputs a parameter at t + 1 (step S1809). After completing the execution of step S1809, the calculation apparatus 101 ends the parameter setting process of the mixed normal distribution. As described above, the parameter setting process of the mixed normal distribution shown in FIG. 18 can generate parameters used in the mixed normal distribution.

  As described above, according to the calculation device, the calculation method, and the calculation program, there is a possibility that the supply source and the supply destination are connected from the amount of change in the electric energy of the distribution board and the outlet at two or more times. An index value indicating is calculated. Thereby, the calculation device can output an index value indicating an estimation result of the connection relation between the distribution board and the outlet without following the wiring.

  Moreover, since the calculation apparatus concerning this Embodiment can estimate a connection relationship, without using the electric power information which a general household appliance consumes, it can perform an estimation process easily. Furthermore, since the calculation apparatus according to the present embodiment does not perform power control of the supply source and the supply destination, the estimation process can be executed at any time. When the change amount at one time is used, the calculation device performs power control of the supply destination and estimates that the supply source change amount and the supply destination change amount are similar to each other. become. However, in the calculation apparatus according to the present embodiment, by using a change amount at two or more times, the possibility that a change amount at the supply source and the change amount at the supply destination are similar is increased. The index value calculated using the amount of change in the time is likely. Therefore, the calculation apparatus according to the present embodiment can calculate an index value indicating a probable connection relationship without performing power control.

  Further, the calculation device generates a probability density function that can be taken by the change amount of the distribution board from the change amount of the electricity amount of the distribution board and the outlet, and obtains the likelihood when connection or disconnection is assumed. Good. The likelihood becomes higher as the assumption is correct. Therefore, the user can grasp the connection relationship by knowing the assumption that the likelihood is high.

  Further, the calculation device may calculate an index value indicating an estimation result of the connection relationship with the likelihood assuming the connection. The higher the likelihood of assuming the connection, the more correct the assumption that the connection is made. Thereby, for example, the user can recognize that the supplier with the highest index value is connected with reference to the output index value.

  Further, the calculation device may calculate an index value indicating an estimation result of the connection relationship with a likelihood assuming no connection. The lower the likelihood assuming no connection, the more incorrect the assumption that no connection is made. Thereby, for example, the user can recognize that the supply source and the supply destination having the lowest index value are connected with reference to the output index value.

  In addition, when obtaining the likelihood assuming connection or the likelihood assuming non-connection, the calculation device uses the number of change data in the distribution board 1 and the outlet 1, the distribution board 2 and the outlet 1. The more accurate index value can be calculated when the number of change amount data matches.

  Further, the calculation device may calculate an index value indicating the possibility that the supply source and the supply destination are connected from the likelihood assumed to be connected and the likelihood assumed not to be connected. The calculation device outputs a result that the supply source having the highest likelihood assumed for connection and the lowest likelihood assuming non-connection is connected to the supply destination. Therefore, the calculation device can output a more accurate result to the user than when one of the likelihood assuming connection or the likelihood assuming non-connection is used. In addition, when the method for obtaining the index value is used, the calculation device is configured such that the number of change data between the distribution board and the outlet 1 and the number of change data between the other distribution board and the outlet 1 are one. Even if not done, an accurate index value can be calculated.

  Further, the calculation device may calculate the likelihood by substituting the ratio of the observed value of the variance of the variation at the supplier and the theoretical value of the variance into the χ square distribution. As a result, the calculation device has a smaller calculation amount than the case of using the probability density function that the change amount of the distribution board can take, and can provide the estimation result to the user earlier. In addition, the calculation device may obtain the likelihood when the connection is assumed using the χ square distribution. Further, the calculation device may obtain the likelihood when non-connection is assumed using the χ square distribution. Further, the calculation device outputs an index value indicating the possibility that the supply source and the supply destination are connected from the likelihood assumed to be connected and the likelihood assumed not to be connected, using the chi-square distribution. Also good.

  In addition, the calculation device may calculate the index value based on the sign of the amount of change between the supply source and the supply destination at two or more times. The calculation device can provide the user with an index value that is not easily influenced by an exceptional change amount even if an exceptional change amount that does not normally occur occurs.

  In addition, based on the index value provided by the calculation device, a building manager who is a user can specify the amount of power used by a department that uses an outlet connected to a distribution board, for example. Using the specified amount of power, for example, a building manager can make a power saving request to the corresponding department.

  The calculation method described in the present embodiment can be realized by executing a program prepared in advance on a computer such as a personal computer or a workstation. This calculation program is recorded on a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, and a DVD, and is executed by being read from the recording medium by the computer. The calculation program may be distributed via a network such as the Internet.

101 Calculation Device 508 Power Amount Change Table 521 Extraction Unit 522 Generation Unit 523 Setting Unit 524 Calculation Unit 525 Calculation Unit 526 Output Unit

Claims (12)

First correspondence information that is time information indicating the amount of change of the measurement value of the measurement object at each supply source of the supply source group that supplies the measurement object and the measurement time at the predetermined time, and the measurement From a storage unit that stores second correspondence information that is time information indicating a change amount of the measurement value of the measurement target at the supply destination to which the target is supplied and the measurement time of the predetermined time The time information of the first correspondence information and the time information of the second correspondence information are the same time and the amount of change at any of the supply sources in the supply source group is the same as the supply destination. An extraction unit for extracting a combination of the change amount of two or more,
Based on a combination of two or more of the change amount at one of the supply sources extracted by the extraction unit and the change amount at the supply destination, the one of the supply sources and the supply destination are connected. A calculation unit for calculating an index value indicating the possibility of being
A calculation device comprising: A generating unit that generates a statistical model indicating a probability distribution that can be taken by the amount of change at any one of the sources based on the amount of change at any of the two or more sources extracted by the extracting unit; ,
Based on at least the amount of change at any one of the sources, the amount of change at any one of the sources is set according to the estimated content indicating the connection or non-connection between any one of the sources and the destination. And a setting unit to
The calculation unit includes:
The change amount set by the setting unit is input to the statistical model generated by the generation unit, and a probability indicating the degree of correctness of the estimated content serving as the index value is calculated. Item 2. The calculation device according to Item 1. The setting unit
When the estimated content indicates a connection, a value obtained by subtracting a change amount at the supply destination from a change amount at any of the supply sources is changed at any of the supply sources according to the estimated content indicating the connection. Set to quantity,
The calculation unit includes:
A change amount corresponding to the estimated content indicating the connection set by the setting unit is input to the statistical model generated by the generating unit, and the degree of correctness of the estimated content indicating the connection serving as the index value is determined. The calculation device according to claim 2, wherein a probability of indicating is calculated. The setting unit
When the estimation content indicates disconnection, the amount of change at any of the supply sources is set to the amount of change at any of the supply sources according to the estimation content indicating disconnection,
The calculation unit includes:
The statistical model generated by the generation unit is input with a change amount according to the estimation content indicating the non-connection set by the setting unit, and the correctness of the estimation content indicating the non-connection as the index value The calculation device according to claim 2, wherein a probability indicating the degree of the calculation is calculated. The setting unit
Furthermore, the amount of change at any one of the suppliers is set to the amount of change at any of the suppliers according to the estimated content indicating disconnection,
The calculation unit includes:
Furthermore, the amount of change corresponding to the estimation content indicating the non-connection set by the setting unit is input to the statistical model generated by the generation unit to indicate the degree of correctness of the estimation content indicating the non-connection. Calculating a probability, and calculating the index value based on a probability indicating a degree of correctness of the estimated content indicating the connection and a probability indicating a degree of correctness of the estimated content indicating the non-connection. The calculation device according to claim 3. The generator is
Depending on the number of change amounts at any one of the two or more supply sources extracted by the extraction unit, a ratio between the observed dispersion value and the theoretical dispersion value at any of the supply sources can be taken. Generate a statistical model showing the probability distribution,
The setting unit
Based on the amount of change at any one of the supply sources and the amount of change at the supply destination, any one of the supplies according to the estimated content indicating connection or non-connection between any of the supply sources and the supply destination Set the observed variance of the original variation and the theoretical variance,
The calculation unit includes:
The statistical model generated by the generating unit is inputted with a ratio between the observed value of variance set by the setting unit and the theoretical value of variance, and indicates the degree of correctness of the estimated content to be the index value The calculation device according to claim 2, wherein a probability is calculated. The setting unit
When the estimated content indicates a connection, the variance of the value obtained by subtracting the change amount at the supply destination from the change amount at any of the supply sources is determined by any of the supply sources according to the estimated content indicating the connection. And a value obtained by subtracting the variance of the amount of change at the supply destination from the variance of the amount of change at any one of the supply sources according to the estimated content indicating the connection Set the theoretical value of the variance of the amount of change at any of the above suppliers,
The calculation unit includes:
The estimated content indicating the connection by inputting a ratio between the observed value of variance and the theoretical value of variance according to the estimated content indicating the connection set by the setting unit to the statistical model generated by the generating unit The calculation device according to claim 6, wherein a probability indicating a degree of correctness of the image is calculated. The setting unit
When the estimated content indicates disconnection, the variance of a value obtained by subtracting the change amount at the supply destination from the change amount at any of the supply sources is any supply according to the estimated content indicating disconnection. Estimated content indicating the non-connection is set to an observed value of the variance of the change amount at the original, and a value obtained by adding the variance of the change amount at the supply destination to the variance of the change amount at any of the supply sources To the theoretical value of the variance of the amount of change at any of the above suppliers according to
The calculation unit includes:
The statistical model generated by the generating unit is input with a ratio between the observed value of variance and the theoretical value of variance according to the estimation content indicating the non-connection set by the setting unit, thereby indicating the non-connection The calculation apparatus according to claim 6, wherein a probability indicating a degree of correctness of the estimated content is calculated. The setting unit
Further, the variance of the value obtained by subtracting the change amount at the supply destination from the change amount at any one of the supply sources is represented by the variance of the change amount at any of the supply sources according to the estimation content indicating disconnection. Any one of the above-mentioned supply according to the estimated content indicating the non-connection is set to an observed value, and a value obtained by adding the variance of the change amount at the supply destination to the variance of the change amount at any of the supply sources Set to the theoretical variance of the original variation,
The calculation unit includes:
Furthermore, the ratio of the observed value of variance corresponding to the estimated content indicating the non-connection set by the setting unit and the theoretical value of variance is input to indicate the degree of correctness of the estimated content indicating the non-connected Calculating a probability, and calculating the index value based on a probability indicating a degree of correctness of the estimated content indicating the connection and a probability indicating a degree of correctness of the estimated content indicating the non-connection. The calculation device according to claim 7. The calculation unit includes:
Calculating the index value based on the sign of the change amount at any one of the two or more supply sources extracted by the extraction unit and the sign of the change amount at the two or more supply destinations. The calculation device according to claim 1, wherein Computer
First correspondence information that is time information indicating the amount of change of the measurement value of the measurement object at each supply source of the supply source group that supplies the measurement object and the measurement time at the predetermined time, and the measurement From a storage unit that stores second correspondence information that is time information indicating a change amount of the measurement value of the measurement target at the supply destination to which the target is supplied and the measurement time of the predetermined time The time information of the first correspondence information and the time information of the second correspondence information are the same time and the amount of change at any of the supply sources in the supply source group is the same as the supply destination. And the change amount of the same time is extracted for two or more,
Based on the extracted change amount at any one of the supply sources at the same time and the change amount at the supply destination at the same time, the supply source and the supply destination are Calculate an index value indicating the possibility of connection,
A calculation method characterized by executing processing. On the computer,
First correspondence information that is time information indicating the amount of change of the measurement value of the measurement object at each supply source of the supply source group that supplies the measurement object and the measurement time at the predetermined time, and the measurement From a storage unit that stores second correspondence information that is time information indicating a change amount of the measurement value of the measurement target at the supply destination to which the target is supplied and the measurement time of the predetermined time The time information of the first correspondence information and the time information of the second correspondence information are the same time and the amount of change at any of the supply sources in the supply source group is the same as the supply destination. And the change amount of the same time is extracted for two or more,
Based on the extracted change amount at any one of the supply sources at the same time and the change amount at the supply destination at the same time, the supply source and the supply destination are Calculate an index value indicating the possibility of connection,
A calculation program for executing a process.

Images