JP2014102779A - Data interpolation device, data interpolation program and data interpolation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To interpolate data with high accuracy.SOLUTION: A power control device 10 includes a time series data model DB 42 for storing one or a plurality of time series models of data and an interpolation method definition DB 46 for each model for storing an interpolation method corresponding to a phase of the time series model, a measured data storage section 22 stores measured data output from a device or a sensor for each device or sensor, and a time series data model determination section 24 determines to which one of the models stored in the time series data model DB a time change pattern of the data stored for each device or sensor corresponds. Also, a phase determination section 28 determines in which one of phases of the model a loss which exists on a part of the data stored is positioned, and an interpolation section 30 interpolate the loss by extracting the interpolation method corresponding to the determined phase from the interpolation method definition DB for each model.

Description

  This case relates to a data interpolation apparatus, a data interpolation program, and a data interpolation method.

  In recent years, for the purpose of reducing power consumption and the like, device control based on measurement data by various sensors and devices connected to a network has been performed. The measurement data in this case is time-series data, but time-series data with no deficiency is necessary to accurately grasp and accurately control the environmental situation and the equipment operation situation.

  Note that the present invention is not limited to the above case, and there are many cases where time-series data having no deficiency is required. Recently, various proposals have been made on methods for interpolating time series data (see, for example, Patent Documents 1 to 6).

JP 2007-33337 A JP 2011-118755 A JP 2011-174737 A JP 2005-222444 A JP-A-8-75219 JP 2002-215646 A

  However, in the technique disclosed in the above publications, interpolation using the same method is executed for time series data of various devices (devices, sensors, etc.). For this reason, there is a possibility that the interpolation accuracy may be lowered depending on the type of device.

  In addition, when interpolating using past similar data as in Patent Documents 5 and 6, similar data must be specified from a large number of data, which may require time and effort.

  In one aspect, an object of the present invention is to provide a data interpolation apparatus, a data interpolation program, and a data interpolation method capable of performing data interpolation with high accuracy.

  The data interpolation device described in this specification specifies a model storage unit that stores one or more time-series models of data and a range in which time changes of data are common in the time-series models stored in the model storage unit An interpolation method storage unit that stores an interpolation method corresponding to a phase to be performed, a data storage unit that stores data output from an external device for each external device, and a time change pattern of data stored for each external device. A model determination unit for determining which one or a plurality of time series models stored in the model storage unit corresponds to, and a deficiency existing in a part of data accumulated for each external device, A phase determination unit that determines in which phase of the time-series model determined by the model determination unit, and a phase determined by the phase determination unit The corresponding interpolation method is extracted from the interpolation method storage unit, and a, interpolation section for interpolating the defect in the interpolation method.

  The data interpolation program described in this specification accumulates data output from an external device for each external device, and the time change pattern of the data accumulated for each external device is predetermined one or more. Determining which of the time series models is applicable, and among the plurality of phases showing a common range of time change of the data of the time series model, missing data existing in a part of the data accumulated for each external device Is a data interpolation program that causes a computer to execute a process of determining a phase in which the position is located and interpolating the defect by an interpolation method according to the determined phase.

  In the data interpolation method described in the present specification, a step of accumulating data output from an external device for each external device and a time change pattern of the data accumulated for each external device are predetermined 1 or A part of data accumulated for each external device among a plurality of phases indicating a range in which a time change of data of the time series model is common and a step of determining which one of the plurality of time series models corresponds to A data interpolation method in which a computer executes a step of determining a phase in which a deficit existing in is located and a step of interpolating the deficit by an interpolation method according to the determined phase.

  The data interpolation device, the data interpolation program, and the data interpolation method described in this embodiment have an effect that data interpolation can be performed with high accuracy.

It is a figure showing roughly the composition of the power control system concerning a 1st embodiment. It is a figure which shows the hardware constitutions of the power control apparatus of FIG. FIG. 2 is a functional block diagram of the power control device in FIG. 1. 4A and 4B are diagrams illustrating an example of data stored in the measurement data DB. It is a figure which shows time series data model DB. It is a figure which shows model DB. It is a figure which shows interpolation method definition DB for every model. FIG. 8A is a flowchart showing the process of the measurement data collection unit, and FIG. 8B is a flowchart showing the process of the measurement data storage unit. It is a flowchart which shows the process of a time series data model determination part. FIG. 10A is a diagram for explaining the base of time-series data and other than the base, and FIG. 10B is a diagram showing a figure representing a line segment other than the base of FIG. 10A. . Fig.11 (a) and FIG.11 (b) are figures (the 1) which show an example of time series data. FIG. 12A and FIG. 12B are diagrams (part 2) illustrating an example of time-series data. It is a flowchart which shows the process of the interpolation availability determination part. FIG. 14A is a flowchart showing processing of the phase determination unit, and FIG. 14B is a flowchart showing processing of the interpolation unit. It is a figure which shows DB every model interpolation method definition DB which concerns on 2nd Embodiment. FIG. 16A is a flowchart showing processing of the interpolation unit according to the second embodiment, and FIG. 16B is a diagram showing an example of data after interpolation. FIG. 17A is a diagram illustrating a continuous interpolation number-reliability table according to the third embodiment, and FIG. 17B is a diagram illustrating an example of data after interpolation. It is a functional block diagram of the electric power control apparatus which concerns on 4th Embodiment. FIG. 19A is a flowchart showing processing of the interpolation unit according to the fourth embodiment, and FIG. 19B is a flowchart showing processing of the inquiry unit according to the fourth embodiment. It is a functional block diagram of the electric power control apparatus which concerns on 5th Embodiment. It is a flowchart which shows the process of the time series data model determination part which concerns on 5th Embodiment. It is a flowchart which shows the process of the new data model production | generation part which concerns on 5th Embodiment. FIG. 23A is a diagram showing an example of a figure newly registered in the non-modeled data DB, and FIGS. 23B and 23C are registered in the non-modeled data DB. It is a figure which shows an example of the figure similar to the figure of Fig.23 (a).

<< First Embodiment >>
Hereinafter, a first embodiment of the power control system will be described in detail with reference to FIGS. FIG. 1 schematically shows the configuration of a power control system 100 according to the first embodiment.

  A power control system 100 illustrated in FIG. 1 includes a device 80 and a sensor 82 as external devices installed in a home, and a power control device 10 as a data interpolation device installed in the home. In the present embodiment, the device 80 and the sensor 82 and the power control device 10 are connected to the home network, so that the power control device 10 and the device 80 and the sensor 82 are connected.

  The device 80 is, for example, a lighting device, an air conditioner, a refrigerator, a television, or the like. When the device 80 has a function of measuring power consumption, air temperature, outside air temperature, humidity, and the like, data obtained by the measurement is output to the power control apparatus 10.

  The sensor 82 is a sensor provided separately from the device 80, and includes a sensor that measures the amount of power consumption in the device 80 that does not have the measurement function, and a sensor that measures temperature, outside temperature, humidity, and the like. Data measured by the sensor 82 is output to the power control apparatus 10. The power control system 100 may include both the device 80 and the sensor 82, but may not include the sensor 82 when the device 80 has the measurement function described above.

  The power control apparatus 10 acquires data output from the device 80 and the sensor 82 and performs control for suppressing power consumption of the device 80 based on the data. In addition, when there is a deficiency in a part of data output from the device 80 or the sensor 82, the power control apparatus 10 interpolates the deficiency and suppresses power consumption based on the complemented data. Take control. Here, when there is a defect in a part of the data and it is necessary to perform interpolation, a part of the data necessary for device control cannot be obtained due to a power failure or the data measurement interval is controlled. It means that it is longer than the required interval and the necessary data is not collected.

  FIG. 2 shows a hardware configuration of the power control apparatus 10. As shown in FIG. 2, the power control apparatus 10 includes a CPU (Central Processing Unit) 90, a ROM (Read Only Memory) 92, a RAM (Random Access Memory) 94, and a storage unit (here, an HDD (Hard Disk Drive)) 96. , An input / output interface 97, a portable storage medium drive 99, and the like. Each component of the power control device 10 is connected to a bus 98. In the power control apparatus 10, a program (including a data interpolation program) stored in the ROM 92 or the HDD 96 or a program (including a data interpolation program) read from the portable storage medium 91 by the portable storage medium drive 99 is stored in the CPU 90. Is executed, the function shown in FIG. 3 is realized.

  FIG. 3 shows a functional block diagram of the power control apparatus 10. As shown in FIG. 3, the power control device 10 includes a measurement data collection unit 20, a measurement data storage unit 22 as a data storage unit, a time series data model determination unit 24 as a model determination unit, and an interpolation availability determination unit. 26, a phase determination unit 28, an interpolation unit 30, a processing unit 32, and a notification unit 34. In FIG. 3, measurement data DB (database) 40 stored in the HDD 96 of the power control apparatus 10, time series data model DB 42 as a model storage unit, model DB 44, and interpolation for each model as an interpolation method storage unit A method definition DB 46 is illustrated.

  The measurement data collection unit 20 collects data (for example, data such as power consumption data, temperature, and room temperature) output from the devices 80 and sensors 82 on the home network and stores the collected data in the measurement data storage unit 22. Send to. In addition, as a data collection method by the measurement data collection unit 20, a method using an IT distribution board, an outlet for measuring electric energy, a sensor data transmission adapter, or the like can be employed. Collection of measurement data for each device 80 and sensor 82 may be performed at regular intervals (for example, every 10 seconds) in each device and sensor, but may be performed at different intervals for each device and sensor. For example, the data collection interval by the temperature sensor with little change may be set every 30 seconds, and the data collection interval by the power amount sensor with much change may be set every 10 seconds. Note that the measurement data collection unit 20 may be provided outside the power control apparatus 10. In this case, the measurement data collection unit 20 may be connected to the home network.

  The measurement data storage unit 22 stores the data transmitted from the measurement data collection unit 20 in the measurement data DB 40 for each device 80 and sensor 82. Thereby, time series data like FIG.4 (a) and FIG.4 (b) is stored in measurement data DB40.

  The time series data model determination unit 24 includes one or a plurality of models (time series models of data) in which the time change pattern of the time series data read from the measurement data DB 40 is stored in the time series data model DB 42 (see FIG. 5). ) Is determined. Then, the time series data model determination unit 24 stores the determination result and the like in the model DB 44 (see FIG. 6).

  Here, the time-series data model DB 42 in FIG. 5 has fields of “model name” and “model shape”. In the “model name” field, “square type”, “mountain type”, “triangular type” and the like are stored as an example. In the “model shape” field, the shape representing the model itself is stored.

  Further, the model DB 44 of FIG. 6 has fields of “device, sensor name”, “model name”, and “parameter”. In the “device, sensor name” field, the names of the device 80 and the sensor 82 for which the time-series data model determination unit 24 performed model determination are stored, and the determined model (corresponding model) is stored in the “model name”. Is stored. In “parameter”, a parameter of time series data for which model determination is performed (extracted by the time series data model determination unit 24) is stored. Details of the parameters will be described later.

  When the time series data read from the measurement data DB 40 has a deficiency, the interpolation availability determination unit 26 refers to the model DB 44 and determines whether or not the deficiency of the read time series data can be interpolated. Here, the interpolability determination unit 26 determines whether or not interpolation is possible based on whether or not the model DB 44 stores the corresponding model of the device or sensor for which the read time-series data is measured. When it is determined that interpolation is possible, the interpolability determination unit 26 notifies the corresponding model to the phase determination unit 28, and when it is determined that interpolation is not possible, the notification unit 34 notifies the fact. Send.

  The phase determination unit 28 refers to the time-series data having a defect and the model-specific interpolation method definition DB 46 (FIG. 7) to determine which phase of the model shape the defect belongs to, and the determination result is an interpolation unit. 30 is notified. Here, “phase” means a range in which time changes are common in each model.

  FIG. 7 shows an interpolation method definition DB 46 for each model. As shown in FIG. 7, the model-by-model interpolation method definition DB 46 has fields of “model name”, “model shape”, “phase”, and “interpolation method”. In the “model name” field, “square type”, “mountain type”, “triangular type” and the like are stored as an example. In the “model shape” field, a shape representing the model itself and a number indicating a range (phase) in which the time change is common in each model are stored. In the “Phase” field, the phase number described in “Model shape” is stored, and in the “Interpolation method” field, interpolation suitable for interpolating the missing data in each phase The contents of the method (interpolation method corresponding to the phase) are stored. Note that, depending on the phase, there may be a case where there is no interpolation method.

  The interpolation unit 30 extracts an interpolation method from the model-specific interpolation method definition DB 46 in FIG. 7 based on the determination result of the phase determination unit 28 (which phase of the corresponding model the defect belongs to). And the interpolation part 30 interpolates a defect | deletion with the extracted interpolation method, and stores the interpolated data in measurement data DB40.

  The processing unit 32 acquires data stored in the measurement data DB 40 (including data obtained by interpolating defects in the interpolation unit 30), and controls the device 80 based on the data (to reduce power consumption). Control).

  The notification unit 34 notifies a terminal used by a user such as a system administrator when the interpolation determination unit 26 or the phase determination unit 28 determines that interpolation cannot be performed.

  Next, data interpolation processing by the power control apparatus 10 of the present embodiment will be described in detail. The data interpolation process is a process executed by the measurement data collection unit 20, the measurement data storage unit 22, the time series data model determination unit 24, the interpolation availability determination unit 26, the phase determination unit 28, and the interpolation unit 30. Hereinafter, processing of each unit will be described.

(Processing of the measurement data collection unit 20)
First, the process of the measurement data collection unit 20 will be described along the flowchart of FIG.

  First, the measurement data collection unit 20 waits until the measurement data collection timing predetermined for each device 80 and sensor 82 is reached in step S10 of FIG. Then, when the measurement data collection timing has come, the measurement data collection unit 20 proceeds to step S12 and collects measurement data. Note that the measurement data is power consumption data measured by the device 80 and the sensor 82.

  In step S <b> 14, the measurement data collection unit 20 transmits the collected measurement data to the measurement data storage unit 22. Thereafter, the process returns to step S10, and the process of FIG.

  In FIG. 8A, data is transmitted to the measurement data storage unit 22 at the timing when the measurement data is collected. However, the present invention is not limited to this. For example, the measurement data collection unit 20 may collect and transmit data to the measurement data storage unit 22 at a stage where a certain amount (predetermined number) of measurement data is collected.

(Processing of measurement data storage unit 22)
Next, the processing of the measurement data storage unit 22 will be described along the flowchart of FIG.

  First, the measurement data storage unit 22 waits until measurement data is transmitted from the measurement data collection unit 20 in step S20 of FIG. That is, in the process of Fig.8 (a), it waits until step S14 is performed. Then, when the measurement data is transmitted from the measurement data collection unit 20, the process proceeds to step S22.

  If transfering it to step S22, the measurement data storage part 22 will store the received measurement data in measurement data DB40, and will return to step S20. Thereafter, the process of FIG. 8B is repeated. Thereby, the measurement data DB 40 stores time-series data for each device and sensor as shown in FIGS. 4 (a) and 4 (b).

(Processing of time-series data model determination unit 24)
Next, the process of the time series data model determination unit 24 will be described with reference to the flowchart of FIG.

  First, the time-series data model determination unit 24 waits until the model determination timing for each device 80 and sensor 82 is reached in step S30 of FIG. As the model determination timing, for example, a timing at which a predetermined time elapses after each of the device 80 and the sensor 82 is connected to the home network can be adopted. When the model determination timing for either the device 80 or the sensor 82 arrives and the determination in step S30 is affirmed, the process proceeds to step S32.

  In step S <b> 32, the time-series data model determination unit 24 extracts data of the device 80 or sensor 82 that has arrived at the model determination timing from the measurement data DB 40.

  Next, in step S34, the time-series data model determination unit 24 extracts time-series data other than the base based on a range where the change is small near the lowest value. For example, in the case of data as shown in FIG. 10A, time series data other than that is extracted based on the range where the value is about 200 (the range where the change is small near the minimum value).

  Next, in step S36, the time series data model determination unit 24 graphs time series data other than the base. Next, in step S38, the time-series data model determination unit 24 represents the graph obtained in step S36 as a line segment. In this case, the line segment expression means connecting portions with large fluctuations with straight lines and connecting the remaining data with straight lines. In this line segment expression, data are connected with as few straight lines as possible, or an approximate curve is generated. Note that FIG. 10B shows a graphic representation of the time series data (other than the base) of FIG.

  Next, in step S40, the time-series data model determination unit 24 determines which of the models included in the time-series data model DB 42 in FIG. In this case, the corresponding model is determined by comparing the graphic representing the line segment with the model included in the time-series data model DB 42 of FIG. In the case of FIG. 10B, the corresponding model is determined to be “square type”.

  Next, in step S42, the time-series data model determination unit 24 determines whether or not there is a corresponding model of a figure represented by a line segment. If the determination is negative, the process returns to step S30. If the determination is positive, the process proceeds to step S44.

  In step S44, the time-series data model determination unit 24 extracts each parameter of the graphic representing the line segment. For example, in the case of time-series data as shown in FIG. 11A, the corresponding model is a square type, and the parameters are “vertical: 1000, horizontal: as shown in the device and sensor name“ A ”in FIG. Undefined is extracted. Further, for example, in the case of time-series data as shown in FIG. 11B, the corresponding model is mountain-shaped, and the parameters are “vertical: 4200, “Horizontal: undefined” is extracted. Further, for example, in the case of time-series data as shown in FIG. 12A, the corresponding model is triangular, and the parameters are “vertical: 100,” as shown in the device and sensor name “C” in FIG. “Horizontal: 10 to 15 M (min), peak position: intermediate, repeat: yes” is extracted. Further, for example, in the case of time series data as shown in FIG. 12B, the corresponding model is triangular, and the parameters are “vertical: 1000,” as shown in the device, sensor name “D” in FIG. “Horizontal: indefinite, peak position: 0 to 1 H (hours), repeat: none” is extracted.

  Returning to FIG. 9, in the next step S <b> 46, the time-series data model determination unit 24 registers the corresponding model and parameter for each device 80 and sensor 82 in the model DB 44. Thereafter, the process returns to step S30, and the process of FIG. 9 is repeated.

  Note that, by repeatedly executing the processing of FIG. 9, information on the corresponding model of the time series data of each device 80 and each sensor 82 is registered in the model DB 44.

  In the above-described processing, the case has been described in which the time series data is represented as a line segment, and then the corresponding model is determined by comparison with the model included in the time series data model DB 42. However, the present invention is not limited to this. . For example, a graph of time series data in a period other than the base and a model included in the time series data model DB 42 may be directly compared to determine the corresponding model. In this case, the model included in the time series data model DB 42 is enlarged or reduced in the vertical and horizontal directions, and the model having the smallest difference from the graph of the time series data is determined as the corresponding model. If the difference does not become smaller than a predetermined value regardless of which model is used, it is determined that there is no corresponding model of the graphic representing the line segment.

(Processing of interpolation enable / disable determining unit 26)
Next, the process of the interpolability determination unit 26 will be described with reference to the flowchart of FIG.

  In the process of FIG. 13, first, in step S50, the interpolability determination unit 26 acquires data of a predetermined range of a predetermined device or a predetermined sensor from the measurement data DB 40. Next, in step S51, the interpolation availability determination unit 26 determines whether there is missing data.

  For example, in the case of a device or sensor that collects data every 10 seconds, if the difference between the collection time of the latest data and the previous data has passed 10 seconds or more, the interpolation availability determination unit 26 Judge that there was data. In addition, for example, if data for every 10 seconds is required for a device or sensor that collects data every 30 seconds, the interpolation feasibility determining unit 26 obtains two pieces of data before and after one collected data. Judge as missing.

  If the determination in step S51 is negative, the process returns to step S50, but if the determination in step S51 is positive, the process proceeds to step S52.

  In step S52, the interpolability determination unit 26 determines whether there is a corresponding model (registered in the model DB 44) in the time series data of the device 80 or the sensor 82 having the missing data. If the determination here is negative, the process proceeds to step S58, and the interpolation availability determination unit 26 notifies the notification unit 34 of data loss. Note that the notification unit 34 notifies the user terminal such as a system administrator that the data is missing but cannot be interpolated at the stage of receiving the notification from the interpolation availability determination unit 26.

  On the other hand, if the determination in step S52 is affirmative, that is, if there is a corresponding model in the time series data of the device or sensor that has missing data, the process proceeds to step S54. In step S54, the interpolability determination unit 26 determines that data interpolation is possible, and the process proceeds to step S56.

  In step S56, the interpolation availability determination unit 26 notifies the phase determination unit 28 of the corresponding model. In addition, after the process of step S58 or step S56 is performed, it returns to step S50 and the interpolation possibility determination part 26 repeats the process of FIG.

(Processing of the phase determination unit 28)
Next, the process of the phase determination part 28 is demonstrated along the flowchart of Fig.14 (a).

  In the process of FIG. 14A, first, in step S60, the phase determination unit 28 waits until the corresponding model is notified from the interpolation availability determination unit 26. When the corresponding model is notified from the interpolation enable / disable determining unit 26 (step S56), the phase determining unit 28 proceeds to step S62 and determines in which phase of the corresponding model the data missing position is. In step S62, the phase determination unit 28 can determine which phase of the corresponding model the data missing position is by confirming a change in data in a predetermined range before and after the data missing position.

  Next, in step S64, the phase determination unit 28 determines whether or not it is an interpolable phase. In this case, if the interpolation method corresponding to the phase determined in step S62 exists in the model-by-model interpolation method definition DB 46 in FIG. 7, the determination in step S64 is affirmed.

  When the determination in step S64 is affirmed, the process proceeds to step S66, and the phase determination unit 28 notifies the interpolation unit 30 of the phase. On the other hand, if the determination in step S64 is negative, the process proceeds to step S68, and the phase determination unit 28 notifies the notification unit 34 of data loss. The notification unit 34 notifies the user terminal such as a system administrator of information that data is missing but cannot be interpolated. In addition, after the process of step S66 or step S68 is performed, it returns to step S60 and the phase determination part 28 repeats the process of Fig.14 (a).

(Processing of the interpolation unit 30)
Next, the process of the interpolation part 30 is demonstrated along the flowchart of FIG.14 (b).

  In the process of FIG. 14B, first, in step S70, the interpolation unit 30 waits until the phase is notified from the phase determination unit 28. When the phase is notified from the phase determination unit 28, the interpolation unit 30 proceeds to step S72, and acquires an interpolation method corresponding to the phase from the model-specific interpolation method definition DB 46. For example, if the corresponding model is “square” and the phase of the data loss position is “2”, the interpolation method “vertical value” is acquired from FIG.

  Next, in step S74, the interpolation unit 30 performs defect interpolation by an interpolation method corresponding to the phase. For example, when “vertical value” is acquired as the interpolation method, the missing data is interpolated with the parameter “vertical value” of the model DB 44.

  As described above, the data in which the defect is interpolated by the interpolation unit 30 is stored in the measurement data DB 40. In addition, after the process of step S74 is performed, it returns to step S70 and the interpolation part 30 repeats the process of FIG.14 (b).

  Note that the processing unit 32 can perform processing (control of the device 80 and the like) using the interpolated data. Thereby, it is possible to perform an appropriate process (appropriate control of the device 80, etc.) compared to a case where a process using non-interpolated data is performed.

  As described above in detail, according to the first embodiment, the power control apparatus 10 stores the time series data model DB 42 (FIG. 5) in which one or more time series models (models) of data are stored, and the model. And a model-specific interpolation method definition DB 46 (FIG. 7) that stores an interpolation method corresponding to a phase specified as a common range in which data changes in time. The measurement data storage unit 22 stores the measurement data output from the device 80 or the sensor 82 for each device or sensor, and the time-series data model determination unit 24 changes the data stored for each device or sensor over time. It is determined whether the pattern corresponds to one or a plurality of models stored in the time series data model DB 42. In addition, the phase determination unit 28 determines in which phase of the corresponding model a defect that exists in a part of the data accumulated for each device 80 and sensor 82, and the interpolation unit 30 is determined. The interpolation method corresponding to the selected phase is extracted from the model-by-model interpolation method definition DB 46, and the defect is interpolated by the interpolation method. Thereby, in the first embodiment, it is possible to interpolate the missing data by an appropriate interpolation method corresponding to the corresponding model of the data in which the defect has occurred and the phase to which the missing position belongs. Therefore, data interpolation can be performed with high accuracy. Moreover, since the process part 32 performs the process based on the data interpolated by the said method, it becomes possible to perform an appropriate process.

  In the first embodiment, the case where the power control apparatus 10 is installed in the home has been described. However, the present invention is not limited to this. For example, at least a part of functions other than the measurement data collection unit 20 may be provided in a server (installed in a data center or the like) connected to a network such as the Internet.

  In the first embodiment, the case where the time series data model DB 42 and the model-by-model interpolation method definition DB 46 are separate databases has been described. However, the present invention is not limited to this. For example, both DBs may be integrated into one DB.

  In the first embodiment, the phase determination unit 28 notifies the interpolation unit 30 of the phase in step S66 of FIG. 14A. However, the present invention is not limited to this. For example, the phase determination unit 28 may notify the interpolation unit 30 of the interpolation method corresponding to the phase.

  In the first embodiment, the power control apparatus 10 may have the functions of the units illustrated in FIG. 3 as a whole. Therefore, some of the functions of a certain function unit (data storage unit, interpolation unit, etc.) in FIG. 3 may be included in another function unit instead of the function unit, or a plurality of function units in FIG. May be combined into one functional unit, or one functional unit may be divided into a plurality of functional units.

<< Second Embodiment >>
Next, the power control system according to the second embodiment will be described in detail with reference to FIGS. 15 and 16.

  In the power control system according to the second embodiment, the model-by-model interpolation method definition DB 46 ′ is partially different from the model-by-model interpolation method definition DB 46, and part of the processing of the interpolation unit 30 is different. ing. Hereinafter, these points will be mainly described.

  FIG. 15 shows a model-by-model interpolation method definition DB 46 ′ according to the second embodiment. In the model-by-model interpolation method definition DB 46 ′ shown in FIG. 15, as can be seen from the model-by-model interpolation method definition DB 46, reliability is set corresponding to each interpolation method. This reliability means that the smaller the value, the more reliable (the higher the reliability). The reliability may be determined by a user such as a system administrator, or automatically set according to the type of phase (base or not, whether the time change is linear or curved, etc.) Good. Note that the reliability value may be modified by the user even if it is set automatically.

  FIG. 16A is a flowchart illustrating processing of the interpolation unit 30 according to the second embodiment. The interpolation unit 30 executes the same processing as in the first embodiment in steps S70 to S74. In step S <b> 76, the interpolation unit 30 assigns reliability to the interpolated data. Specifically, the interpolation unit 30 extracts the reliability from the model-specific interpolation method definition DB 46 ′ based on the phase of the interpolated data, and adds reliability to the interpolated data as shown in FIG. Is granted.

  As described above, by assigning reliability corresponding to the phase to the interpolated data, the processing unit 32 can perform processing based on the reliability. For example, if the threshold value of reliability is set to 2 in the processing unit 32, the processing unit 32 does not use interpolation data whose reliability value is larger than the threshold value = 2 in the process, and does not use the reliability data. It is possible to handle such as using only interpolation data whose value is smaller than threshold = 2. As described above, the processing unit 32 performs processing in consideration of the reliability, so that it is possible to realize appropriate processing with high accuracy.

<< Third Embodiment >>
Next, the power control system according to the third embodiment will be described in detail with reference to FIG.

  In the power control system according to the third embodiment, instead of the model-by-model interpolation method definition DB 46 ′ of the second embodiment, the model-by-model interpolation method definition DB 46 and the reliability shown in FIG. A continuous interpolation number-reliability table 148 as a storage unit is used.

  FIG. 17A shows a continuous interpolation number-reliability table 148 according to the third embodiment. As shown in FIG. 17A, each field has “continuous interpolation number” and “reliability”, and the number of continuous interpolation of data and the reliability corresponding to the number are defined. . In FIG. 17A, the smaller the reliability value, the more reliable (the higher the reliability) is. That is, in FIG. 17A, it is defined that the smaller the number of continuous interpolations, the higher the interpolation accuracy and the higher the reliability.

  In the third embodiment, the same processing as in the second embodiment is performed. That is, the interpolation unit 30 executes processing according to the flowchart of FIG. However, in the case where the interpolating unit 30 assigns reliability to the interpolated data in step S76, the interpolating unit 30 sets the reliability corresponding to the number of consecutive interpolated data (number of continuous interpolations) to the continuous interpolation number-reliability table 148 ( It is extracted from FIG. As a result, as shown in FIG. 17B, reliability corresponding to the number of continuous interpolations can be given to each interpolated data.

  As described above, the interpolating unit 30 assigns reliability corresponding to the number of continuous interpolations to the interpolated data, so that the processing unit 32 can perform processing based on the reliability. For example, if the threshold value of reliability is set to 2 in the processing unit 32, the processing unit 32 does not use interpolation data whose reliability value is larger than the threshold value = 2 in the process, and does not use the reliability data. It is possible to handle such as using only interpolation data whose value is smaller than threshold = 2. As described above, the processing unit 32 performs processing in consideration of the reliability, so that it is possible to realize appropriate processing with high accuracy.

  Note that the reliability described in the second and third embodiments may be used. For example, if the interpolated data has a reliability level of 2 according to the phase and a reliability level of 3 according to the number of continuous interpolations, the reliability is the sum of 2 and 3 (= 5) or the product of 2 and 3 (= 6 ). Moreover, it is good also as adding all the reliability of the time series data of a certain range, or calculating the average value etc., and making these values the reliability of the whole time series data of the said range. .

<< Fourth Embodiment >>
Next, the power control system according to the fourth embodiment will be described in detail with reference to FIGS.

  The power control system according to the fourth embodiment includes a power control apparatus 10 'shown in FIG. The power control apparatus 10 ′ has an inquiry unit 36 as can be understood by comparing FIG. 18 and FIG. 3. As the model-by-model interpolation method definition DB, the model-by-model interpolation method definition DB 46 ′ of FIG. 15 may be used, as in the second embodiment, or, as in the third embodiment, model-by-model interpolation in FIG. The method definition DB 46 may be used together with the continuous interpolation number-reliability table 148.

  Hereinafter, the processing of the interpolation unit 30 and the inquiry unit 36 in the fourth embodiment will be described with reference to the flowcharts of FIGS. 19A and 19B.

  The interpolation unit 30 stands by until the phase is notified from the phase determination unit 28 in step S170 of FIG. When the phase is notified from the phase determination unit 28, the interpolation unit 30 proceeds to step S172, and acquires an interpolation method corresponding to the phase from the model-by-model interpolation method definition DB 46 '(or 46). Further, the interpolation unit 30 acquires the reliability corresponding to the phase or the reliability corresponding to the continuous interpolation number from the model-by-model interpolation method definition DB 46 ′ or the continuous interpolation number-reliability table 148.

  Next, in step S173, it is determined whether or not the acquired reliability is greater than or equal to a predetermined threshold value. When judgment here is denied, it transfers to step S180. In step S180, the interpolation unit 30 performs interpolation using the acquired interpolation method, assigns reliability to the interpolated data, and returns to step S170. On the other hand, if the determination in step S173 is affirmed, the process proceeds to step S174.

  In step S174, the interpolation unit 30 notifies the inquiry unit 36 of completion of interpolation preparation together with the reliability.

  On the other hand, the inquiry unit 36 is on standby until it receives a notification from the interpolation unit 30 in step S90 of FIG. 19B, and after the processing of step S174 is performed, the process proceeds to step S92.

  In step S92, the inquiry unit 36 transmits an inquiry screen to a terminal of a user such as a system administrator. It is assumed that the reliability value is also displayed on this inquiry screen. Next, in step S94, it is determined whether or not the user's terminal has received an NG answer (an answer indicating that interpolation should not be performed) from the user. If the determination here is affirmative, the process proceeds to step S100, NG (not to be interpolated) is answered to the interpolation unit 30, and the process returns to step S90. On the other hand, if the determination in step S94 is negative, the process proceeds to step S95, and it is determined whether or not an OK answer (answer that interpolation can be performed) is received from the user at the user terminal. If the determination here is affirmative, the process proceeds to step S98, the interpolating unit 30 is answered OK (interpolation may be performed), and the process returns to step S90. On the other hand, if the determination in step S95 is negative, the process proceeds to step S96, where an inquiry screen is transmitted to the user's terminal, and it is determined whether a preset waiting time has elapsed. . If the determination is negative, that is, if the waiting time has not elapsed, the process returns to step S94. On the other hand, if the determination in step S96 is affirmative, that is, if the waiting time has elapsed, it is considered that the user terminal has answered OK, and the process proceeds to step S98. After step S98, the process returns to step S90.

  By the way, when a response is transmitted from the inquiry unit 36 to the interpolation unit 30 in step S98 or S100, the determination in step S176 (whether the response is received from the inquiry unit 36) in FIG. 19A is affirmed, and step S178 is performed. Migrate to In step S178, the interpolation unit 30 determines whether interpolation is possible, that is, whether the answer from the user is OK. If the determination is negative, the process returns to step S170 without performing data interpolation. If the determination is positive, the process proceeds to step S180. When the process proceeds to step S180, the interpolation unit 30 performs data interpolation using the interpolation method acquired in step S172, and assigns reliability to the interpolated data.

  In the above description, the case where it is considered that the user has answered OK when there is no response from the user has been described. However, the present invention is not limited to this.

  As described above, according to the fourth embodiment, since the inquiry unit 36 asks the user whether or not to perform interpolation, the user can be assured that data that should not be interpolated is not interpolated. A feeling can be obtained. In addition, since the user is inquired whether or not to perform interpolation only when the reliability is greater than the threshold value (when the reliability is not so high), the number of inquiries to the user can be reduced and efficient. Processing is possible.

  In the fourth embodiment, the inquiry unit 36 inquires the user whether or not to perform interpolation when the reliability is equal to or greater than the threshold value. However, the present invention is not limited to this. That is, the inquiry unit 36 may inquire the user whether or not to perform interpolation for all interpolations regardless of the reliability.

<< Fifth Embodiment >>
Next, the power control system according to the fifth embodiment will be described in detail with reference to FIGS.

  The power control system according to the fifth embodiment includes a power control apparatus 10 ″ shown in FIG. 20. As can be seen from a comparison between FIG. 20 and FIG. An unmodeled data DB 48 and a new data model generation unit 38 are included.

  The non-modeled data DB 48 is a database that stores a graphic representing a line segment of the time-series data when the time-series data model determination unit 24 determines that there is no corresponding model corresponding to the predetermined time-series data. It is.

  The new data model generation unit 38 generates a new model based on a plurality of line segment representations stored in the non-modeled data DB 48 and registers it in the model DB 44.

  Next, the processing of the time series data model determination unit 24 and the processing of the new data model generation unit 38 will be described based on FIGS. 21 and 22.

  FIG. 21 is a flowchart showing the processing of the time-series data model determination unit 24 in the fifth embodiment. In the process of FIG. 21, as compared with the process of the first embodiment (the process of FIG. 9), when the determination of step S <b> 42 (whether there was a corresponding model) is denied, step S <b> 48 is executed, The process returns to step S30.

  In step S48, the time-series data model determination unit 24 registers the graphic representing the line segment in the non-modeled data DB 48. Note that the non-modeled data DB 48 is registered with graphics representing line segments as shown in FIGS. 23 (a) to 23 (c).

  FIG. 22 is a flowchart showing the processing of the new data model generation unit 38. In the process of FIG. 22, first, in step S110, the new data model generation unit 38 waits until a new figure is registered in the non-modeled data DB. In this case, when the time series data model determination unit 24 executes the process of step S48 of FIG. 21, the process proceeds to step S112.

  In step S112, the new data model generation unit 38 determines whether or not there are a predetermined number or more of figures similar to the new figure in the non-modeled data DB 48. For example, when the predetermined number is 2, the figure shown in FIG. 23A is newly registered, and the figures of FIG. 23B and FIG. 23C are stored in the non-modeling data DB 48. The determination is affirmed and the process proceeds to step S114. On the other hand, if the determination in step S112 is negative, the process returns to step S110. Whether or not the figures are similar is determined by whether or not the figures to be compared are enlarged or reduced in the vertical and horizontal directions and the difference (area) between the two figures is smaller than a predetermined area. .

  In step S114, the new data model generation unit 38 extracts a new figure and a figure with the smallest number of line segments from similar figures. In the case of FIGS. 23A to 23C, the figure of FIG. 23A having the smallest number of line segments is extracted. In the fifth embodiment, the extracted figure becomes a new model.

  Next, in step S116, the new data model generation unit 38 generates parameters. The parameters generate vertical, horizontal, peak positions, repeats, etc. based on the change points of the line segments in the figure, and each line segment is determined as a phase. The phase of the base portion is 0.

  Next, in step S118, the new data model generation unit 38 deletes a new figure and a figure similar to the new figure (the figure contributing to the new model generation) from the non-modeled data DB 48. Next, in step S120, the new data model generation unit 38 registers the figure and parameters represented by the line segment in the model DB 44, and returns to step S110. The model registered in the model DB 44 may be registered in the time series data model DB 42 or the model-specific interpolation method definition DB 46 by a user such as a system administrator, for example. Alternatively, the new data model generation unit 38 may be automatically registered in the time series data model DB 42 and the model-specific interpolation method definition DB 46. When the new data model generation unit 38 automatically registers a new model, the interpolation method depends on the phase (line segment) shape (straight line or curved line, how much the inclination is, etc.). It may be determined.

  In addition, after returning to step S110, the said process is repeatedly performed.

  As described above in detail, according to the fifth embodiment, the time-series data model determination unit 24 has a corresponding model of time-series data (a graphic representing a line segment) accumulated for each device / sensor. If not, time series data (figure representing line segment) is accumulated in the non-modeled data DB 48, and the new data model generation unit 38 selects a plurality of similar time series data (line) from the accumulated time series data. A figure representing a segment) is extracted, and a new model is generated based on the extracted time-series data (a figure representing a line segment). Thereby, since a new model can be generated based on time-series data, in the case where data interpolation cannot be performed in the first to fourth embodiments only by using a model prepared in advance, Interpolation using a new model can be performed.

  In the fifth embodiment, the graphic representing the line segment in step S48 of FIG. 21 is registered in the non-modeled data DB 48. However, the present invention is not limited to this, and the time series data itself may be registered. Good. In this case, when a predetermined number or more of similar time-series data exists in the non-modeled data DB 48, each time-series data may be represented as a line segment, and a line-segment representation figure having the smallest number of line segments may be used as a new model. .

  In each of the above embodiments, the case where the power control apparatus 10 collects data from the home device 80 and the sensor 82 has been described. However, the present invention is not limited to this. For example, the power control apparatus 10 may collect data from a device outside the house (for example, a sensor that measures outside air temperature or a device that an external organization (such as the Japan Meteorological Agency) has) via a network such as the Internet.

  In each of the above embodiments, the case where the power control apparatus 10 has functions of the measurement data storage unit 22, the time series data model determination unit 24, the interpolation availability determination unit 26, the phase determination unit 28, the interpolation unit 30, and the like will be described. However, it is not limited to this. That is, as long as it is an apparatus that needs to interpolate data, the function of each unit may be provided in another apparatus.

  The above processing functions can be realized by a computer. In that case, a program describing the processing contents of the functions that the processing apparatus should have is provided. By executing the program on a computer, the above processing functions are realized on the computer. The program describing the processing contents can be recorded on a computer-readable recording medium (except for a carrier wave).

  When the program is distributed, for example, it is sold in the form of a portable recording medium such as a DVD (Digital Versatile Disc) or a CD-ROM (Compact Disc Read Only Memory) on which the program is recorded. It is also possible to store the program in a storage device of a server computer and transfer the program from the server computer to another computer via a network.

  The computer that executes the program stores, for example, the program recorded on the portable recording medium or the program transferred from the server computer in its own storage device. Then, the computer reads the program from its own storage device and executes processing according to the program. The computer can also read the program directly from the portable recording medium and execute processing according to the program. Further, each time the program is transferred from the server computer, the computer can sequentially execute processing according to the received program.

  Each embodiment mentioned above is an example of suitable implementation of the present invention. However, the present invention is not limited to this, and various modifications can be made without departing from the scope of the present invention.

In addition, regarding the above description of the first to fifth embodiments, the following additional notes are disclosed.
(Supplementary Note 1) A model storage unit that stores one or more time series models of data;
An interpolation method storage unit that stores an interpolation method corresponding to a phase that is specified as a common range of time variation of data in the time series model stored in the model storage unit;
A data storage unit for storing data output from the external device for each external device;
A model determination unit that determines whether a time change pattern of data accumulated for each external device corresponds to one or a plurality of time-series models stored in the model storage unit;
A phase determination unit that determines in which phase of the time-series model determined by the model determination unit a defect that exists in a part of the data accumulated for each external device; and
A data interpolation device comprising: an interpolation unit that extracts an interpolation method according to the phase determined by the phase determination unit from the interpolation method storage unit and interpolates the deficiency by the interpolation method.
(Additional remark 2) The said interpolation method storage part stores the reliability of the said interpolation method with the interpolation method corresponding to the said phase,
2. The data interpolation according to claim 1, wherein the interpolation unit acquires a reliability corresponding to the phase in which the defect exists from the interpolation method storage unit, and gives the reliability to the data after the interpolation. apparatus.
(Additional remark 3) It further has the reliability storage part which stores the reliability according to the continuous number of the said defect | deletion,
The interpolation unit specifies the number of consecutive missing data, acquires a reliability corresponding to the specified number from the reliability storage unit, and adds the reliability to the data after interpolation. The data interpolation device according to Supplementary Note 1 or 2, which is a feature.
(Additional remark 4) The data interpolation apparatus of Additional remark 2 or 3 further provided with the inquiry part which inquires whether a user performs the said interpolation.
(Supplementary note 5) The data interpolating apparatus according to supplementary note 4, wherein the inquiry unit makes the inquiry when a reliability given to the interpolated data is lower than a predetermined reliability. .
(Supplementary Note 6) When the model determination unit determines that the time change pattern of data accumulated for each external device does not correspond to any time-series model stored in the model storage unit, An accumulator for accumulating time change patterns;
A new data model generation unit that extracts a plurality of similar time change patterns from the time change patterns stored in the storage unit and generates a new model based on the extracted time change patterns; The data interpolation device according to any one of Supplementary notes 1 to 4.
(Appendix 7) Accumulating data output from external devices for each external device,
Determining whether a time change pattern of data stored for each external device corresponds to one or a plurality of time-series models determined in advance;
Of a plurality of phases indicating a time range of data of the time series model indicating a common range, determine a phase in which a defect that exists in a part of the data accumulated for each external device is located,
A data interpolation program characterized by causing a computer to execute a process of interpolating the defect by an interpolation method according to the determined phase.
(Supplementary note 8) The data interpolation program according to supplementary note 7, wherein in the interpolating process, a reliability corresponding to the phase in which the defect exists is added to the data after the interpolation.
(Additional remark 9) In the said process to interpolate, the number which the said defect | deletion continued was specified, and the reliability according to the said specified number is provided to the data after the said interpolation, The additional remark 7 or 8 characterized by the above-mentioned. Data interpolation program.
(Additional remark 10) The data interpolation program of Additional remark 8 or 9 which makes the said computer further perform the process which inquires a user whether the said interpolation is performed.
(Supplementary note 11) The data interpolation program according to supplementary note 10, wherein the inquiry process is executed when a reliability given to the interpolated data is lower than a predetermined reliability.
(Supplementary Note 12) When it is determined that the time change pattern of the data stored for each external device does not correspond to any time series model, the time change pattern is stored,
A plurality of similar time change patterns are extracted from the accumulated time change patterns, and a new model is generated based on the extracted time change patterns, further causing the computer to execute a process. The data interpolation program according to any one of appendices 7 to 11.
(Additional remark 13) The process which accumulate | stores the data output from an external device for every external device,
Determining whether a time change pattern of data accumulated for each external device corresponds to one or a plurality of time-series models determined in advance;
Determining a phase in which a defect that exists in a part of the data accumulated for each of the external devices is located among a plurality of phases indicating a common range of time change of data of the time series model;
A data interpolation method, wherein the computer executes a step of interpolating the deficiency by an interpolation method corresponding to the determined phase.
(Supplementary note 14) The data interpolation method according to supplementary note 13, wherein in the step of interpolating, a reliability corresponding to a phase in which the defect exists is added to the data after the interpolation.
(Supplementary note 15) The supplementary note 13 or 14, wherein, in the step of interpolating, the number of consecutive defects is specified, and a reliability corresponding to the specified number is given to the data after the interpolation. Data interpolation method.
(Supplementary note 16) The data interpolation method according to supplementary note 14 or 15, wherein the computer further executes a step of inquiring a user whether or not to perform the interpolation.
(Supplementary note 17) The data interpolation method according to supplementary note 16, wherein the inquiring step is executed when a reliability given to the interpolated data is lower than a predetermined reliability.
(Supplementary note 18) When it is determined that the time change pattern of data stored for each external device does not correspond to any time series model, the step of storing the time change pattern;
A step of extracting a plurality of similar time change patterns from the accumulated time change patterns and generating a new model based on the extracted time change patterns; 18. The data interpolation method according to any one of appendices 13 to 17, which is characterized.

10 Power control device (data interpolation device)
22 Measurement data storage unit (data storage unit)
24 Time-series data model determination unit (model determination unit)
28 Phase determination unit 30 Interpolation unit 36 Inquiry unit 38 New data model generation unit 42 Time series data model DB (model storage unit)
46 Interpolation method definition DB for each model (interpolation method storage)
48 Non-modeled data DB (storage unit)
80 equipment (external equipment)
82 Sensor (external device)
148 Continuous Interpolation Number-Reliability Table (Reliability Storage Unit)

Claims (7)

A model storage unit for storing one or more time series models of data;
An interpolation method storage unit that stores an interpolation method corresponding to a phase that is specified as a common range of time variation of data in the time series model stored in the model storage unit;
A data storage unit for storing data output from the external device for each external device;
A model determination unit that determines whether a time change pattern of data accumulated for each external device corresponds to one or a plurality of time-series models stored in the model storage unit;
A phase determination unit that determines in which phase of the time-series model determined by the model determination unit a defect that exists in a part of the data accumulated for each external device; and
A data interpolation device comprising: an interpolation unit that extracts an interpolation method according to the phase determined by the phase determination unit from the interpolation method storage unit and interpolates the deficiency by the interpolation method. The interpolation method storage unit stores the reliability of the interpolation method together with the interpolation method corresponding to the phase,
2. The data according to claim 1, wherein the interpolation unit acquires a reliability corresponding to a phase in which the deficiency exists from the interpolation method storage unit, and gives the reliability to the data after the interpolation. Interpolator. A reliability storage unit that stores the reliability according to the number of consecutive defects;
The interpolation unit specifies the number of consecutive missing data, acquires a reliability corresponding to the specified number from the reliability storage unit, and adds the reliability to the data after interpolation. The data interpolation device according to claim 1 or 2, characterized in that   The data interpolation device according to claim 2, further comprising an inquiry unit that inquires of a user whether to perform the interpolation. When the model determination unit determines that the time change pattern of the data accumulated for each external device does not correspond to any time series model stored in the model storage unit, the time change pattern is An accumulating unit for accumulating;
A new data model generation unit that extracts a plurality of similar time change patterns from the time change patterns stored in the storage unit and generates a new model based on the extracted time change patterns; The data interpolation device according to any one of claims 1 to 4. Accumulate data output from external devices for each external device,
Determining whether a time change pattern of data stored for each external device corresponds to one or a plurality of time-series models determined in advance;
Of a plurality of phases indicating a time range of data of the time series model indicating a common range, determine a phase in which a defect that exists in a part of the data accumulated for each external device is located,
A data interpolation program characterized by causing a computer to execute a process of interpolating the defect by an interpolation method according to the determined phase. Storing data output from the external device for each external device;
Determining whether a time change pattern of data accumulated for each external device corresponds to one or a plurality of time-series models determined in advance;
Determining a phase in which a defect that exists in a part of the data accumulated for each of the external devices is located among a plurality of phases indicating a common range of time change of data of the time series model;
A data interpolation method, wherein the computer executes a step of interpolating the deficiency by an interpolation method corresponding to the determined phase.

Images