JP2013009500A - Power management device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a power management device for precisely grasping an operational state of an electric apparatus.SOLUTION: Management means 5 stores a waveform of operational current originated from a power supply circuit of an electric apparatus as a unique waveform; while acquisition means 1 and 2 acquire waveforms of current supplied to the electric apparatus as power supply waveforms, and extraction means 4 extracts difference waveforms among periods of the supply power waveform. Estimation means 6 estimates an apparatus class of the electric apparatus 14 whose operational state has changed, by comparing the unique waveform stored by the management means 5 and the difference waveforms extracted by the extraction means 4.

Description

  The present case relates to a power management apparatus that estimates an operating state of an electric device that operates by receiving power supply.

  Conventionally, for the purpose of power management in offices, factories, houses, etc., power management devices that monitor the operating states of a plurality of electrical devices have been used. As a technique for identifying an electric device in operation, an identification method based on the amount of power consumed by the electric device or a variation pattern of power consumption is known. For example, the power consumption pattern consumed by each electrical device is stored in a database as a template in advance, and the actual power consumption is measured based on the current value measured at the power line inlet. A typical method is to identify an operating electric device by comparing a power pattern with a template (see Patent Document 1).

  Also, spectrum analysis (dissolving current value for each frequency component and calculating the intensity of each frequency component) is performed on the current time-series data, and the electrical equipment in operation is determined from the spectrum distribution state. A method of identifying is known (see Patent Document 2). Alternatively, based on the intensity ratio of the harmonics in the current change derived from the inverter circuit of the electrical equipment, a method for estimating the current change (operating state) for each model using the independent component analysis method is also known ( (See Patent Document 3). By using such a method, it is said that each operation state can be grasped while distinguishing electric devices having different power consumption patterns and power consumption amounts.

JP 2005-354794 A Japanese Patent No. 3922358 Japanese Patent No. 4454001

  However, the identification method based on the power consumption pattern and its spectrum requires a comparison operation for all combinations of electrical devices to be monitored. On the other hand, the number of combinations of electrical devices increases exponentially according to the number and type of electrical devices. Therefore, even in cases such as offices and factories where a large number of electrical devices exist, or in ordinary housing cases where there are various types of electrical devices, There may be a case where the operating state of the electrical equipment cannot be grasped.

  On the other hand, in the method using independent component analysis, it is not necessary to prepare a model specific to each electrical device in advance, and it is possible to identify current changes of each electrical device that operates autonomously and asynchronously. . However, this technique is based on the premise that the harmonic intensity ratio of each device is unique to the model of the electrical device, and there is a problem that it is difficult to distinguish between models with low linear independence. is there.

  For example, when the current change of a certain model is close to the scalar multiple of the current change of another model, it is difficult to distinguish whether the former is operating or whether the latter is operating a plurality. In the case where the number of models obtained as a result of analysis is given as a real number, the solution may be a decimal number, and the practical problem that the number of actually operating units becomes indistinguishable Is also present.

One of the purposes of this case has been created in view of these problems, and is to accurately grasp the operating state of electrical equipment.
In addition, the present invention is not limited to the above-described object, and is an operational effect derived from each configuration shown in “Mode for Carrying Out the Invention” to be described later. It can be positioned as a purpose.

  The disclosed power management apparatus includes a management unit that stores a waveform of an operating current derived from a power circuit of an electric device as a specific waveform, an acquisition unit that acquires a waveform of a supply current to the electric device as a power supply waveform, and the power supply Extracting means for extracting a differential waveform between waveform periods. And an estimation unit configured to estimate a model of the electrical device whose operating state has changed by comparing the natural waveform with the differential waveform.

According to the disclosed technology, at least one of the following effects and advantages can be obtained.
(1) The estimation accuracy of the model of the electrical equipment can be improved.
(2) It is possible to grasp a change in the operating state of the electrical equipment.
(3) It is possible to identify a specific waveform of an electric device having poor linear independence.

It is a block diagram which shows the whole structure of the power management apparatus which concerns on embodiment. It is a figure for demonstrating the detection target by a power management apparatus, (a) is a graph which illustrates the waveform of an electric current, (b) is a table | surface which illustrates the data which sampled this waveform. (A) is a graph illustrating a power supply waveform recorded in time series, and (b) is a table illustrating a data group obtained by sampling the power supply waveform. It is a figure for demonstrating the calculation which concerns on the extraction of the difference between the periods of a feeding waveform, (a) is a graph which shows the calculation object of a difference, (b) is a graph which showed the magnitude | size of the difference visually, (c) ) Is a graph illustrating a differential waveform. It is a figure for demonstrating the calculation which concerns on the extraction of the difference between periods of an electric power feeding waveform, (a) is a graph which shows the calculation object of a difference, (b) is a graph which showed the magnitude | size of the difference visually. It is a graph which illustrates the model-specific current data which a power management apparatus memorize | stores, (a) illustrates the specific waveform when the electric equipment of each model is turned on, (b) is turned off This is an example of the natural waveform of time. It is a flowchart which illustrates the control content implemented with an electric power management apparatus. It is a flowchart for demonstrating the modification of the control implemented with an electric power management apparatus. It is a flowchart for demonstrating the modification of the control implemented with an electric power management apparatus. It is a graph for demonstrating the power management apparatus which concerns on a modification, and sets the specific waveform which a power management apparatus memorize | stores for every state of a model. It is a graph for demonstrating the power management apparatus which concerns on the modified example which made the waveform target part collation.

  Embodiments according to the present power management apparatus will be described below with reference to the drawings. However, the embodiment described below is merely an example, and there is no intention to exclude various modifications and technical applications that are not clearly shown in the embodiment described below. In other words, the present embodiment can be implemented with various modifications (combining the embodiments and modifications) without departing from the spirit of the present embodiment.

[1. Device configuration]
FIG. 1 illustrates a power supply system in which the power management apparatus 10 is provided in the vicinity of a power supply line inlet of a commercial AC power supply. Here, the power management apparatus 10 is interposed on the feeder line 11 connected to the distribution board 13. The distribution board 13 is provided in a building 12 of a power consumer such as a general house, an office building, or a factory. Inside the distribution board 13, the feeder line 11 is branched and formed into a plurality of lines, and each line is wired to various places in the building 12. Various electric devices 14 are connected to each line.

  For example, in the case of an office building, examples of the type of the electric device 14 include lighting devices such as fluorescent lamps and incandescent lamps, air conditioners, and electronic computers (computers). In the case of a general house, home appliances such as a refrigerator, a washing machine, a television receiver, and a rice cooker are also included. The power management apparatus 10 individually identifies a wide variety of electrical devices 14 and estimates the respective operating states.

  Regarding the identification of the electric equipment 14, the power management apparatus 10 identifies the change in the power supply state based on the characteristic waveform of the operating current derived from the structure of the power supply circuit built in each electric equipment 14, and the state By accumulating changes, the operating states of all the electric devices 14 in the building 12 are grasped.

  The characteristic waveform is a circuit of a power supply circuit of the electric device 14 when the electric device 14 is supplied with a predetermined alternating voltage (typically a sine wave) and the operating state of the electric device 14 changes from off to on. It means a waveform of a current value (that is, a waveform indicating the amount of change) that varies depending on the structural characteristics and the electrical characteristics of components (for example, capacitors, coils, etc.) interposed on the circuit. The characteristics of these power supply circuits affect the waveform of the current value and the phase of the current value with respect to the voltage value.

  For example, different types of devices such as an electronic computer (computer), an air conditioner, and a lighting device have different current waveforms when the same AC voltage is applied, that is, have different inherent waveforms. In addition, even if the same computer is different, if the manufacturer or product number is different, the specific waveforms are different unless the characteristics of the power supply circuit are the same. Furthermore, even if the same manufacturer and the same product number are used, if the parts used in the power supply circuit are changed (for example, revisions are updated, etc.), the specific waveform changes accordingly.

  Therefore, the specific waveform can be regarded as a waveform specific to the power supply circuit of the electrical device 14. The power management apparatus 10 uses such a property to identify the electrical device 14 whose operating state has changed. Hereinafter, each category in which the electric device 14 is classified according to the shape of the specific waveform is referred to as a “model”. FIG. 1 shows a power feeding system in which the number of models of the electrical equipment 14 fed from the distribution board 13 is m (arbitrary natural number).

[2. Control configuration]
The power management apparatus 10 performs control for grasping the operating state of the electrical device 14, and for example, an LSI (Large Scale) integrated with a microprocessor, ROM (Read Only Memory), RAM (Random Access Memory), and the like. Integration) may be manufactured as a device, or may be a system constructed with a general-purpose electronic computer. As shown in FIG. 1, the power management apparatus 10 includes a current measurement unit 1, a current waveform data acquisition unit 2, a recording unit 3, a difference extraction unit 4, a model-specific current data management unit 5, a model estimation unit 6, and an operating device. A counting unit 7 and an output unit 8 are provided.

  The functions of these current measurement unit 1, current waveform data acquisition unit 2, recording unit 3, difference extraction unit 4, model-specific current data management unit 5, model estimation unit 6, operating device counting unit 7 and output unit 8 are: It may be realized by an electronic circuit (hardware), may be programmed as software, or a part of these functions may be provided as hardware and the other part may be software. Good.

[2-1. Current measurement unit]
The current measuring unit 1 measures a current value of a supply current (total current) supplied from the feeder line 11. The “current value of the supply current” here is not an effective value used for calculation of a general electric energy but an instantaneous value. Further, in the case where the feeder line 11 is a multi-phase, it is sufficient if it measures at least one of the current values. For example, when the power supply line 11 is a three-phase three-wire AC power line, the current measuring unit 1 measures an instantaneous value of a current value of one of three phases.

  The frequency at which the current measuring unit 1 samples the current value (number of measurements per unit time) is set sufficiently higher than the AC frequency of the supplied current (for example, 50 [Hz]), for example, about 20 [kHz]. . It is preferable that at least the sampling frequency is higher than twice the AC frequency of the supply current. The measured current value data is transmitted to the current waveform data acquisition unit 2.

[2-2. Current waveform data acquisition unit]
The current waveform data acquisition unit 2 (acquisition means) arranges the current value data measured by the current measurement unit 1 in time series, and generates a data group of the current value _{I_0} for one cycle. Here, if the AC frequency of the supply current is set to f ₁ [Hz] and the sampling frequency in the current measuring unit 1 is set to f ₂ [Hz], the current waveform data acquisition unit 2 sets the current value data to f ₂ / Hz. classified into _one each f, and classified each data group is transmitted to the recording unit 3.

  When the AC frequency of the supply current is 50 [Hz], the time (cycle) corresponding to one wavelength of the waveform of the current value is 20 [ms]. Here, when the time when the current value is 0 and the change gradient is positive is t = 0, the fluctuation graph for one period of the current value is from time 0 to 20 [ms as shown in FIG. It is expressed within the range up to]. At this time, if the sampling frequency in the current measuring unit 1 is 20 [kHz] (that is, a cycle of 0.05 [ms], and the sampling frequency is 20 [times / ms]), the current value of 20 [ms] The data is measured 400 times, and the coordinates of the sampled measurement points are 400 point coordinates located on the graph.

  Therefore, the data transmitted to the recording unit 3 is a data group including 400 times t and current values corresponding to the power supply waveform for one wavelength, as shown in FIG. As described above, the current waveform data acquisition unit 2 has a function of acquiring a waveform of one cycle of the supply current as a power supply waveform.

[2-3. Recording section]
The recording unit 3 records the data group transmitted from the current waveform data acquisition unit 2 in the storage device 3a in order of time. Specific examples of the storage device 3a include a register and a main memory. The storage device 3a has a storage capacity for storing a data group for a predetermined period (for example, n period, n is an arbitrary natural number). Here, as shown in FIG. 3 (a), a number # 0 is assigned to a data group for one cycle of the latest supply current, and a number #n is assigned to a data group of n cycles in the past, The recording unit 3 defines the data group # 0 as a data group _{I_0} (t) whose argument is the time t within one period of the current value, and records this in the storage device 3a. In other words, the recording unit 3 records the current value in the storage device 3a as a function of discretely given time t.

Further, when the data group _{I_0} (t) is recorded in the storage device 3a, if the data group _{I_0} (t) already exists (that is, the data group that is one cycle earlier is the previous calculation cycle). when that has been recorded) redefines recorded data group I _{_0} a (t) as one cycle before the data group I _{_1} (t), and records it in the storage unit 3a. Similarly, if there are old data groups _{I_1} (t) to I_ _(n-1) (t), they are sequentially lowered to _{I_2} (t) to _{I_n} (t), and the storage device 3a. To record. Therefore, as shown in FIG. 3B, the storage device 3a always holds a history of changes in the power supply waveform up to n cycles in the past.

[2-4. Difference extraction unit]
The difference extraction unit 4 (extraction means) extracts a difference between cycles of the power supply waveform as a difference waveform. Here, as shown in FIG. 4A, the difference between the past power supply waveform and the latest power supply waveform is calculated. For example, the difference x _{_i} (t) between the data group I _{_0} (t) of the latest power supply waveform and the data group I _{_i} (t) of the power supply waveform before i periods (i is a natural number less than n) is expressed by the following equation: Given by 1.

When the sampling frequency in the current measuring unit 1 is 20 [kHz], the data group I _{_i} (t) of each feeding waveform includes data of 400 current values, and accordingly, the difference x _{_i} (t) also includes 400 pieces of difference information.

Here, FIG. 4B shows a graph of the latest power supply waveform data group _{I_0} (t), the power supply waveform data group _{I_i} (t) before i cycles, and the difference _{x_i} (t). . When the area of the region surrounded by each graph and the time axis t is S ₁ , S ₂ , S ₃ , the area S ₃ is equal to the value obtained by subtracting the area S ₂ from the area S ₁ . That is, the graph of the difference x _{_i} (t) is as shown if the current value of the data group I _{_0} (t) for the data group I _{_i} (t) is the extent to which increased or decreased at each time t.

Hereinafter, the waveform represented by the difference _{x_i} (t) calculated here is referred to as a difference waveform. Difference extraction unit 4, such operation is repeated for past each feeding waveform, calculates the n-number of the difference _{x _1 (t) ~x _n (} t). The differential waveform extracted through this calculation is illustrated in FIG.

  The difference waveform has a flat shape whose value is almost zero when the operating state of the electric device 14 is not changed. On the other hand, when the operating state of any of the electrical devices 14 changes, the shape of the differential waveform also changes before and after the change of the operating state. In addition, regarding the electrical equipment 14 of any model, the shape of the differential waveform that occurs when the operating state changes from on to off is the same as the shape of the differential waveform that occurs when the operating state changes from off to on. The shape is reversed.

  FIG. 5A illustrates another power supply waveform different from that shown in FIG. Since the difference extraction unit 4 calculates the difference between the past power supply waveforms with respect to the latest power supply waveform, the shape of the difference waveform obtained by the difference extraction unit 4 depends on the history of the power supply waveform that has fluctuated up to that point. Will do.

For example, when the detection order of the latest power supply waveform and the power supply waveform before i cycles is opposite to that shown in FIG. 4B, as shown in FIG. 5B, the difference _{x_i} (t) This graph is inverted with respect to the time axis t. Note that a differential waveform extracted from the power supply waveform of FIG. 5A is illustrated in FIG. Thus, the differential waveform has a shape that depends on the power supply waveform and its detection order.

[2-5. Current data management section by model]
The model-specific current data management unit 5 (management means) stores the waveform of the operating current corresponding to the power supply circuit of the electrical device 14 as a specific waveform. Specific examples of the model-specific current data management unit 5 include a nonvolatile memory and a storage. Here, for example, as shown in FIG. 6A, for each of the m types of models, a data group y _{_1} (t) to y for one cycle of the natural waveform generated when the power supply is changed from OFF to ON. _{_m} (t) is stored in advance. These data groups _{y _1 (t) ~y _m (} t) is used for matching with the model estimating section 6.

  If the AC voltage supplied to each model is the same, the specific waveform is unchanged unless the power supply circuit changes. Therefore, not only models that receive power supply from the distribution board 13 but also data related to specific waveforms of models that can be newly connected are stored in the model-specific current data management unit 5 so that the power supply system can be expanded. Easy to handle.

In addition, the method for acquiring data related to the specific waveform is arbitrary. For example, a method of extracting a change in the power supply waveform when the power source of only the target model is changed from OFF to ON on the downstream side of the distribution board 13 can be considered. . That is, the intrinsic waveform includes a waveform of the supply current detected by the current measurement unit 1 when the power source is off and a waveform of the supply current detected by the current measurement unit 1 when the power source is on. It can be acquired as a differential waveform. In this case, each of the data groups y _{_1} (t) to y _{_m} (t) of eigen waveforms includes data of 400 current values, and the estimation calculation by the model estimation unit 6 described later becomes easy. .

  Alternatively, the shape of the natural waveform may be calculated theoretically based on the structure of the power supply circuit. In the power management apparatus 10, since the waveforms are collated through an operation for digitizing the similarity between the waveforms, the number and density (sampling frequency) of data constituting each waveform are not necessarily the same. It does not have to be.

[2-6. Model estimation section]
Model estimating unit 6 (estimating means), m pieces of each of the n difference x _{_1} calculated by the difference extraction section _{4 (t) ~x _n (t} ), stored in the model-specific current data management unit 5 Are collated with each of the data groups y _{_1} (t) to y _{_m} (t) of eigen waveforms.

As a means for collating them, the model estimation unit 6 calculates the similarity between the difference waveform extracted by the difference extraction unit 4 and the unique waveform stored in the model-specific current data management unit 5. For example, based on the difference x _{_i} (t) based on the power supply waveform before i cycles (i is a natural number less than n) and the data group y _{_j} (t) of the characteristic waveform of the model j (j is a natural number less than m) The first index _{e_ij} is calculated based on the following equation 2.

The first index _{e_ij} is a sum of squares (sum of squares) of a difference current corresponding to a waveform difference between the difference waveform extracted by the difference extraction unit 4 and the specific waveform stored in the model-specific current data management unit 5. The higher the similarity between the two waveforms (the more similar the shape), the closer to 0. Therefore, when the first index _{e_ij} is smaller than the predetermined threshold δ (δ> 0), the model estimation unit 6 estimates that the two waveforms “match”, and the first index _{e_ij} is _equal to or greater than the threshold δ. In the case of, it is estimated that the two waveforms are “mismatched”.

The coincidence between the differential waveform based on the power supply waveform i cycles before and the unique waveform of model j means that the power state of model j is off at the time before i cycles and is on at the current cycle. To do. For example, if the difference x _{_i} (t) shown in FIG. 4B matches the model 2 unique waveform data group y _{_2} (t) in FIG. _6A , the power state of the model 2 is turned off. Will be changed from on to off.

Further, the model estimation unit 6 calculates the similarity between the difference waveform obtained by reversing the sign of the characteristic waveform with respect to the time axis t. For example, based on the difference x _{_i} (t) based on the power supply waveform i cycles before and the data group y _{_m} (t) of the characteristic waveform of the j-th model, the second index e _{_ijR} is _calculated based on the following equation 3. Calculate. Note that a data group _{y_m} (t) of the characteristic waveform of the model j inverted with respect to the time axis t is expressed as _{-y_m} (t). As shown in FIG. 6B, the graph of the inverted natural waveform coincides with the inversion of the sign of the current value in the graph shown in FIG.

The second index _{e_ijR} is a sum of squares (sum of squares) of the difference current corresponding to the waveform difference between the unique waveform and the difference waveform whose signs are inverted, and the higher the similarity between the two waveforms (the more similar the shape is) The closer it is to 0). Model estimating section 6, when the second index e _{_IjR} is smaller than the threshold value [delta], and estimated that the two waveforms are "match", when the second index e _{_IjR} is equal to or greater than the threshold [delta], two waveforms Is a “mismatch”.

The coincidence between the difference waveform based on the power supply waveform before i cycle and the sign of the inherent waveform of model j reversed is “the power state of model j is on at the time before i cycle, and at the time of the current cycle. It means "off." For example, FIG. 5 (b) and the difference x _{_i} shown in (t), if the the 6 data group specific waveform model 2 (b) in -y _{_2} (t) matches, the power state of the model 2 It is changed from on to off.
The estimation result in the model estimation unit 6 is transmitted to the operating device counting unit 7.

[2-7. Operating equipment counting unit]
The operating device counting unit 7 counts the operating state of the electrical device 14 based on the estimation result in the model estimating unit 6. Here, the number of operating units for each of the m types of models is counted.

First, when it is estimated that two waveforms match by estimation using the first index _{e_ij} , the operating device counting unit 7 increases the operating number of the model j by one. On the other hand, when it is estimated that the two waveforms match by the estimation using the second index _{e_ijR} , the operating device counting unit 7 decreases the operating number of the model j by one. In any case of these operating device count unit 7 transmits a deletion signal to clear all of the power supply waveform I _{_1} used in the estimation of the (t) ~I _{_n} (t) to the recording unit 3.

In addition, when it is estimated that the two waveforms do not match in the estimation using the first index _{e_ij} and the second index _{e_ijR} , the operating device counting unit 7 counts the number of operating models of each model currently counted. Is kept as it is. Information on the number of operating units of each model counted here is transmitted to the output unit 8.

[2-8. Output section]
The output unit 8 outputs the number of operating units of each model counted by the operating device counting unit 7. Here, for example, in response to a request from the user, information such as the number of operating models 1 to m and the time when the operating number has changed is presented on the display device 8a. Note that information on the number of operating units of each model may be output to the outside of the power management apparatus 10 via a communication line or a network (not shown).

[3. flowchart]
As an example of a control procedure executed by the power management apparatus 10, a flowchart is shown in FIG. In step A10, the count in the operating unit counting unit 7 is initialized, and the operating units from model 1 to model m are reset to zero. At this time, it is assumed that the power sources of all the electrical devices 14 connected to the distribution board 13 are off. In Step A20, the current value detected by the current measuring unit 1 is transmitted to the current waveform data acquiring unit 2, and a time-series data group of the current value _{I_0} for one cycle is generated.

In step A30, the data group obtained in the previous step is defined by the recording unit 3 as a data group _{I_0} (t) having the time t as an argument, and recorded in the storage device 3a. For example, when the data group _{I_0} (t) has been recorded in the storage device 3a in the previous calculation cycle, the recorded data group is redefined as the data group one cycle before and stored together. Recorded in the device 3a.

In step A40, the difference extraction unit 4 extracts a difference waveform between cycles of the power supply waveform. Here, according to Equation 1 above, the difference x _{_i} the data group I _{_i} data group I _{_0} (t) and i cycles before the power supply waveform of the latest feeding waveform (t) (t) is calculated. The number of differences _{x_i} (t) calculated here depends on the number of periods n of the past power supply waveforms stored in the storage device 3a.

In step A50, the model estimation unit 6 reads the data group _{y_j} (t) for one period of the unique waveform stored in the model-specific current data management unit 5. The number of data groups _{y_j} (t) read here depends on the number m of models of the electrical equipment 14.

In step A60, the model estimation unit 6 collates n differential waveforms with m natural waveforms. Here, based on the difference _{x_i} (t) and the data group _{y_j} (t), the first index _{e_ij} is calculated according to Equation 2 above. Note that the number of first indices _{e_ij} calculated here is n × m.

In step A70, the model estimation unit 6 selects the smallest one of all the first indices _{e_ij} . In other words, here, the combination that is estimated to have the highest similarity by selecting the difference waveform and the specific waveform is selected. In Step A80, it is determined whether or not the first index _{e_ij} selected in the previous step is less than the first threshold δ.

Here, when the first index _{e_ij} <the first threshold value δ, the process proceeds to step A90, where the operating device counting unit 7 counts that the operating number of the model j has increased by one, and the process proceeds to step A140. In step A <b> 140, the updated operating number of each model is output from the output unit 8. In addition, a deletion signal is transmitted from the operating device counting unit 7 to the recording unit 3, and information on all the power supply waveforms _{I_1} (t) to _{I_n} (t) recorded in the storage device 3a is deleted. Thereafter, the control proceeds to step A20, and the estimation control of the operating state of the electric device 14 is continued.

On the other hand, if it is determined in step A80 that the first index e — _ij ≧ first threshold value δ, the process proceeds to step A100. In step A100, the model estimation unit 6 collates n differential waveforms with m inverted eigen waveforms. Here, based on the difference _{x_i} (t) and the data group _{-y_j} (t), the second index _{e_ijR} is calculated according to Equation 3 above. Note that the number of second indexes _{e_ijR} calculated here is n × m, similarly to the first index _{e_ij} .

In step A110, as in the control in step A70, the smallest one of all the second indices _{e_ijR} is selected. In subsequent step A120, it is determined whether or not the second index _{e_ijR} selected in the previous step is less than the first threshold value δ.

Here, if the second index _{e_ijR} <the first threshold value δ, the process proceeds to step A130, where the operating device counting unit 7 counts that the operating number of the model j has decreased by one, and the process proceeds to step A140. If the second index _{e_ijR} ≧ first threshold value δ in step A120, it is assumed that there is no change in the operating state of the electric device 14, and the process proceeds to step A20 as it is.

[4. effect]
As described above, in the power management apparatus 10 described above, the waveform of the feeding current in the vicinity of the feeding line inlet to the distribution board 13 to which various types of electrical devices 14 are connected is acquired, and the difference waveform between the periods. Is extracted. On the other hand, the waveform of the operating current of the electrical device 14 is stored as a specific waveform, and the differential waveform and the specific waveform are collated to estimate a model whose operating state has changed from off to on or from on to off.

  Therefore, by estimating the operation target based on the characteristic waveform of the operation current derived from the structure of the power supply circuit built in the electric device 14, the characteristic waveform with low linear independence can be easily identified. The estimation accuracy of the model can be improved. That is, even when the current change during operation of a certain model is close to a scalar multiple of the current change during operation of another model, the power management apparatus 10 can identify these models as long as the unique waveforms are different. can do. In addition, by using the differential waveform of the feeding current as a waveform comparison target, it is possible to accurately grasp the change in the operating state of the electric device 14 as the change in the supply current.

  Further, in the power management apparatus 10 described above, the time width for checking the current waveform is set to a width corresponding to one cycle of the alternating current, that is, the specific waveform of one wavelength and the differential waveform of one wavelength are checked. This makes it possible to accurately estimate the model in a short time while reducing resources related to the estimation of the model of the electrical device 14.

  Further, in the power management apparatus 10 described above, the waveforms are collated with a time width corresponding to one cycle of the AC current. The change in current value is the basis for model estimation. Therefore, the necessity of assuming a situation in which the operating states of a plurality of models change at the same time is extremely low, and changes in the operating state of the electrical device 14 can be determined individually.

  Further, in the power management apparatus 10 described above, when the similarity between the unique waveform and the differential waveform is high, it is estimated that the electric device 14 of a model corresponding to the unique waveform has been operated. On the other hand, when the similarity between the inverted natural waveform and the differential waveform is high, it is estimated that the electrical device 14 of that model has stopped operating. As described above, by using the eigen waveform and the inverted eigen waveform obtained by inverting the eigen waveform with respect to the time axis, not only the model of the electric device 14 can be specified but also the operation and stop thereof can be accurately grasped. In addition, since the number of operating models is given as a positive integer, it is possible to avoid a situation in which the number of actually operating models cannot be determined.

Furthermore, in the power management apparatus 10 described above, when determining the degree of similarity between the unique waveform and the difference waveform, the sum of squares of the difference current corresponding to the waveform difference is set as the first index _{e_ij} and is compared with the first threshold value δ. Thus, it is determined whether or not the two waveforms match. Thus, by using the sum of squares of the difference currents, it is possible to improve the calculation speed while improving the estimation accuracy.

The same applies to the determination of the similarity between the inverted intrinsic waveform and the difference waveform, and the sum of squares of the difference current corresponding to the waveform difference is set as the second index _{e_ijR,} and this is compared with the first threshold value δ. It is determined whether two waveforms match. Thus, by using the sum of squares of the difference currents, it is possible to improve the calculation speed while improving the estimation accuracy. Furthermore, since the threshold value δ related to the determination of the first index _{e_ij} and the threshold value δ related to the determination of the second index _{e_ijR} are the same, the condition determination can be performed with the same logic, and the control configuration is simplified. There is an advantage that is easy.

[5. Modified example]
Regardless of an example of the disclosed embodiment, various modifications can be made without departing from the spirit of the present embodiment. Each structure and each process of this embodiment can be selected as needed, or may be combined suitably. In the following modifications, the same elements as those in the above-described embodiment are denoted by the same reference numerals, and description thereof is omitted.

[5-1. First modification]
In the power management apparatus 10, as shown in FIG. 3A, a history of changes in the power supply waveform up to n cycles before is retained, and history information is not cleared unless the operating state of the electrical device 14 changes. That is, when the operating state of the electric device 14 does not change for a long time, there is a possibility that a huge storage capacity is required for the storage device 3a. Therefore, if the newly acquired power supply waveform of the supply current is substantially the same as the previously acquired power supply waveform, it is determined that the operating state of the electrical device 14 has not changed, and the power supply waveform is discarded. A new power supply waveform may be acquired in the next cycle.

As an example of the control procedure in this case, a flowchart is shown in FIG. This flowchart is executed between steps A20 to A40 in the flow of FIG.
In step B10, the data group obtained in step A20 is defined by the recording unit 3 as a data group _{I_0} (t) with time t as an argument. At this time, the data group _{I_0} (t) is not yet recorded in the storage device 3a, and the data that has been recorded in the previous calculation cycle is maintained without being changed.

In step B20, the data group _{I_1} (t) of the previous cycle recorded in the storage device 3a is read into the recording unit 3. In step B30, the recording unit 3 collates the latest power supply waveform with the power supply waveform of the previous cycle. Here, two types of power supply waveform data group I _{_0} (t), based on the I _{_1} (t), the third index e _{_01} is calculated according to equation 4 below.

The third index _{e_01} becomes a value closer to 0 as the similarity between the latest power supply waveform and the power supply waveform of the previous cycle is higher (as the shape is similar). Therefore, the recording unit 3 estimates that the two waveforms “match” when the third index _{e_01} is smaller than a predetermined second threshold θ (θ> 0).

That is, in step B40, the recording unit 3 determines whether or not the third index _{e_01} is less than the second threshold value θ. Here, when the third index _{e_01} <the second threshold value θ, it is determined that the latest power supply waveform and the power supply waveform of the previous cycle are the same, and there is no electrical device 14 whose operating state has changed. Then, the process proceeds to step B50 to discard the latest power supply waveform data group _{I_0} (t). Control thereafter proceeds to step A20 in the flow of FIG.

On the other hand, if the third index e — ₀₁ ≧ second threshold θ is satisfied in Step B40, the process proceeds to Step B60, and the data group I — ₀ (t) is recorded in the storage device 3a. Further, when the data group _{I_0} (t) has been recorded in the storage device 3a in the previous calculation cycle, the recorded data group is redefined as the data group of the previous cycle, and is also stored in the storage device. Recorded in 3a. Thereafter, the control proceeds to step A40 in the flow of FIG.

  In this control, the difference waveform is discarded when no change is seen in the power supply waveform, and the verification in the model estimation unit 5 is not performed. Thereby, the estimation calculation in a state where the operating state of the electric device 14 is not changed can be prevented, and the estimation accuracy of the model of the electric device 14 can be improved. In addition, it is possible to reduce both resources for holding data related to verification and verification costs.

Further, in this control, when determining a state in which no change is found in the power supply waveform, the sum of squares of the difference current corresponding to the waveform difference between the power supply waveforms is set as the third index _{e_01} and compared with the second threshold value θ. Thus, it is determined whether or not the two waveforms match. As described above, by using the sum of squares of the difference currents, it is possible to improve the estimation accuracy with respect to the state of the power supply waveform, to further improve the estimation accuracy of the operating state of the electric device 14, and to reduce the calculation speed. Can be improved.

  In the first modification, when the same power supply waveform is detected in succession, one of the power supply waveforms detected later is discarded. You may change a control aspect suitably. For example, when the same power supply waveform is detected in succession, one of the previously detected power supply waveforms may be discarded, or when three or more same power supply waveforms are detected, the last It is good also as a structure which discards the electric power feeding waveform detected by (1).

[5-2. Second modification]
In the power management apparatus 10 described above, there may be a case where random fluctuations unrelated to the operating state of the electrical device 14 are mixed in the waveform of the current value measured by the current measuring unit 1. For example, when noise is generated on the power supply line 11 side or the power supply voltage fluctuates slightly, the power supply waveform fluctuates even when the operating state of the electrical device 14 is maintained constant. Therefore, it is possible to collate the power supply waveform of an arbitrary period with the power supply waveform of the period before and after that, and exclude the power supply waveform mixed with sudden random fluctuations from the verification target in the model estimation unit 6. .

As an example of the control procedure in this case, a flowchart is shown in FIG. This flowchart is executed between steps A20 to A40 or A50 in the flow of FIG.
In step C10, the data group obtained in step A20 is defined by the recording unit 3 as a data group _{I_0} (t) with time t as an argument. Further, when the data group _{I_0} (t) has been recorded in the storage device 3a in the previous calculation cycle, the recorded data group is redefined as the data group of the previous cycle, and is also stored in the storage device. Recorded in 3a. When both the first modification and the second modification described above are performed, the control procedure may proceed to step C20 of this flow after step B60 in FIG.

At step C20, data group I _{_0} all feeding waveforms stored in the storage unit _{3a (t) ~I _n (t} ) is read in the recording unit 3. In subsequent step C30, the recording unit 3 collates the power supply waveform before k cycles with the power supply waveform before k + 1 cycles for the natural number k defined in the range of n−1 or less.

For example, the fourth index _{a_k} is calculated according to the following equation 5 based on the data groups _{I_k + 1} (t) and _{I_k} (t) of two types of power supply waveforms. Furthermore, in step C40, the power supply waveform before the k-1 cycle is compared with the power supply waveform before the k cycle. For example, the fifth index _{b_k} is calculated according to the following equation 6 based on the data groups _{I_k} (t) and _{I_k-1} (t) of two types of power supply waveforms.

In the subsequent step C50, it is determined whether or not the absolute value | _{a_k- b_k} | of the difference between the fourth index _{a_k} and the fifth index _{b_k} is larger than a predetermined third threshold value α. That is, here, the similarity of the power supply waveform from k + 1 cycles to k cycles before represented by the fourth index _{a_k} and the power supply from k cycles to k−1 cycles represented by the fifth index _{b_k.} It is determined how much the difference from the waveform similarity is.

In step C50, if the absolute value of the difference | _{a_k- b_k} |> the third threshold value α, the process proceeds to step C60, and it is considered that random fluctuations are mixed in the power supply waveform before k cycles. , A control signal for excluding the power supply waveform before k cycles from the extraction target of the differential waveform _{x_i} (t) is transmitted from the recording unit 3 to the differential extraction unit 4. On the other hand, when the absolute value of the difference | _{a_k− b_k} | ≦ the third threshold value α is determined in Step C50, it is determined that random fluctuations are not mixed in the power supply waveform before k cycles, and Step C70. Proceed to

In step C70, for all the power supply waveform from the previous cycle of the power supply waveform to the n-1 previous cycle of the power supply waveform, the absolute value of the difference between the fourth index a _{_k} a fifth index b _{_k} | a _{_k} -b _{_k} It is determined whether or not | Here, when all the inspections are not completed, the value of k is changed and the process returns to Step C30, and whether or not random fluctuations are mixed is repeatedly confirmed. On the other hand, when the 100% inspection is completed, the subsequent control proceeds to step A40 in FIG.

  In this control, the power supply waveform when random fluctuation is mixed in the power supply change is excluded from the extraction target of the differential waveform, and collation in the model estimation unit 5 is not performed. Thereby, the erroneous detection in the state with no change in the operating state of the electric equipment 14 can be prevented, and the estimation accuracy of the model of the electric equipment 14 can be improved.

Further, in the control of the second modified example, when determining whether or not random fluctuations are mixed, the square sum of the difference current corresponding to the waveform difference from the power supply waveform in the preceding and following periods is calculated as the fourth index _{a_k} and the fifth index b. _{_k} is used, and the absolute value of these differences is compared with the third threshold value α to determine whether there is a sudden change. Thus, by using the sum of squares of the difference currents, it is possible to further improve the estimation accuracy of the operating state of the electrical device 14 while accurately eliminating the influence of noise and disturbance that may be included in the power supply waveform.

Note that the method of detecting random fluctuations is not limited to this. For example, when the absolute value of each of the fourth index a _{_k} and the fifth index b _{_k} is equal to or greater than a predetermined value, it may be determined that random fluctuations are mixed, or power supply before k cycles A differential waveform from the waveform to the power supply waveform before k + 1 cycles and a differential waveform from the power supply waveform before k-1 cycles to the power supply waveform before k cycles are calculated, and these shapes are almost mutually on the time axis t. If the shape is inverted, it may be determined that random fluctuations are mixed.

[5-3. Others]
Regarding the specific waveform recorded in the model-specific current data management unit 5, in the above-described embodiment, data corresponding to the model of the electrical device 14 connected to the downstream side of the distribution board 13 as shown in FIG. It has been exemplified that the group _{y _1 (t) ~y _m (} t) is stored in advance, and the model and a unique waveform of the electrical equipment 14 data may not correspond necessarily one to one.

  For example, the electrical device 14 having a difference in operation mode such as strong, medium, and weak may have a different specific waveform depending on the operation state. In this case, as shown in FIG. 10, a data group corresponding to each mode may be set for a model having a plurality of operation modes. As described above, by preparing a data group of specific waveforms for each operating state in advance for each model, not only the operating state (ON / OFF) of the electrical device 14 but also its operating mode can be accurately estimated. Can do.

  Further, in the above-described embodiment, the model estimation unit 6 exemplifies the one that collates the difference waveform and the unique waveform for one period, but the waveform collation technique and the collation range are not limited to this. For example, it is good also as control which collates the difference waveform and intrinsic waveform for two periods or more, and it is good also as control which collates the difference waveform and intrinsic waveform for a half period conversely.

In addition, the collation range can be further narrowed depending on the specific waveform of the model to be estimated. For example, when the number of models to be estimated is three and the characteristic waveforms for the models A, B, and C have shapes as shown in FIG. 11, at least a section T in which the shapes of the characteristic waveforms greatly differ is collated. If so, these models can be identified. Therefore, a section T from time t ₁ to time t ₂ may be set as a verification target section, and a power supply waveform detection section and a differential waveform extraction section may be set in accordance with the section T.

As a method for setting the section T, for example, an analysis of variance considering the model as a factor for the current value of each model at each time t is performed, and the time t (or section) where the difference is dominant is selected. Can do.
Such section settings can further reduce the computational load and resources related to the estimation of the model of the electric device 14, thereby enabling accurate model estimation in a short time.

  Further, in the above-described embodiment, the waveform comparison is exemplified mainly using the sum of squares of the difference currents as an index of similarity. However, a specific method of grasping the similarity of the waveform is arbitrary. The absolute value of the current may be used as an index of similarity, and the sum of the squares of the difference current may be used as an index.

[6. Addendum]
The following supplementary notes are further disclosed with respect to the above embodiments and modifications.

(Appendix 1)
Management means for storing the waveform of the operating current derived from the power supply circuit of the electrical device as a specific waveform;
Obtaining means for obtaining a waveform of a current supplied to the electrical device as a power supply waveform;
Extracting means for extracting a difference waveform between cycles of the power supply waveform;
An electric power management apparatus comprising: an estimation unit configured to estimate a model of the electric device whose operating state has changed by comparing the natural waveform with the differential waveform.

(Appendix 2)
The management means stores the characteristic waveform for one cycle of the operating current,
The acquisition means acquires the power supply waveform for a plurality of cycles of the supply current,
The power management apparatus according to appendix 1, wherein the extraction unit extracts the differential waveform for at least one period from the power supply waveforms for the plurality of periods.

(Appendix 3)
The estimating means is
When the similarity between the natural waveform and the differential waveform is high, it is estimated that the electrical device of the model corresponding to the natural waveform has been operated,
It is estimated that the electrical device of the model corresponding to the specific waveform has stopped operating when the similarity between the inverted specific waveform obtained by inverting the specific waveform with respect to the time axis and the differential waveform is high. The power management apparatus according to appendix 1 or 2.

(Appendix 4)
The estimation means determines that the similarity is high when a sum of squares of a difference current corresponding to a waveform difference between the characteristic waveform and the difference waveform is less than a first threshold. 3. The power management apparatus according to 3.

(Appendix 5)
Any one of Supplementary notes 1 to 4, wherein the extraction unit discards one of the power supply waveforms when the similarity between the pair of the power supply waveforms related to the extraction of the difference waveform is high. The power management device according to item.

(Appendix 6)
The extraction means determines that the similarity between the power supply waveforms is high when the sum of squares of the difference currents corresponding to the waveform difference between the power supply waveforms is less than a second threshold, The power management apparatus according to appendix 5.

(Appendix 7)
Any one of appendices 1 to 6, wherein the extraction unit excludes the power supply waveform including random fluctuations from the extraction target of the differential waveform based on the similarity between the power supply waveforms that are temporally changed. The power management apparatus according to item 1.

(Appendix 8)
The extraction means for any of the power supply waveforms,
The sum of squares of the difference current corresponding to the waveform difference from the immediately preceding feeding waveform is calculated as the first sum of squares, and
While calculating the sum of squares of the difference current corresponding to the waveform difference from the power supply waveform immediately after as the second sum of squares,
The power management apparatus according to appendix 7, wherein when the difference between the first sum of squares and the second sum of squares exceeds a third threshold, it is determined that the random fluctuation is included.

(Appendix 9)
The management means, for the electrical equipment having a plurality of operating modes, stores the unique waveform that differs for each operating mode,
The power management apparatus according to any one of appendices 1 to 8, wherein the estimation unit estimates the operation mode in addition to a model of the electrical device.

(Appendix 10)
The management means stores the characteristic waveform having a predetermined characteristic section which is at least a half wavelength or less and is distinguished from the operating current of the other electrical equipment,
The power management apparatus according to any one of appendices 1 to 9, wherein the estimation unit collates the characteristic section of the characteristic waveform with a section corresponding to the characteristic waveform.

(Appendix 11)
Store the waveform of the operating current derived from the power supply circuit of the electrical equipment as a specific waveform,
Obtain the waveform of the current supplied to the electrical equipment as a power supply waveform,
Extracting the difference waveform between the periods of the power supply waveform,
A power management method characterized by estimating a model of the electrical device in which the operating state has changed by comparing the characteristic waveform with the differential waveform.

(Appendix 12)
Storing the characteristic waveform for one cycle of the operating current;
While obtaining the power supply waveform for a plurality of cycles of the supply current,
The power management method according to appendix 11, wherein the differential waveform for at least one cycle is extracted from the power supply waveforms for the plurality of cycles.

(Appendix 13)
When the similarity between the natural waveform and the differential waveform is high, it is estimated that the electrical device of the model corresponding to the natural waveform has been operated,
It is estimated that the electrical device of the model corresponding to the specific waveform has stopped operating when the similarity between the inverted specific waveform obtained by inverting the specific waveform with respect to the time axis and the differential waveform is high. The power management method according to Supplementary Note 11 or 12.

(Appendix 14)
14. The power management according to appendix 13, wherein the similarity is determined to be high when a sum of squares of difference currents corresponding to a waveform difference between the characteristic waveform and the difference waveform is less than a first threshold value. Method.

(Appendix 15)
The power according to any one of appendices 11 to 14, wherein when one of the pair of power supply waveforms related to the extraction of the differential waveform has a high degree of similarity, any one of the power supply waveforms is discarded. Management method.

(Appendix 16)
The power according to claim 15, wherein when the sum of squares of the difference currents corresponding to the waveform difference between the power supply waveforms is less than a second threshold, it is determined that the similarity between the power supply waveforms is high. Management method.

(Appendix 17)
The supplementary waveform according to any one of appendices 11 to 16, wherein the power supply waveform including random fluctuations is excluded from the extraction target of the differential waveform based on the similarity between the power supply waveforms that move back and forth in time. Power management method.

(Appendix 18)
For any said power supply waveform,
The sum of squares of the difference current corresponding to the waveform difference from the immediately preceding feeding waveform is calculated as the first sum of squares, and
While calculating the sum of squares of the difference current corresponding to the waveform difference from the power supply waveform immediately after as the second sum of squares,
18. The power management method according to appendix 17, wherein when the difference between the first sum of squares and the second sum of squares exceeds a third threshold value, the random change is determined to be included.

(Appendix 19)
For the electrical equipment having a plurality of operating modes, store the unique waveform different for each operating mode,
The power management method according to any one of appendices 11 to 18, wherein the operation mode is estimated in addition to a model of the electrical device.

(Appendix 20)
Storing the characteristic waveform having a predetermined characteristic section which is at least a half-wavelength section or less and is distinguished from the operating current of the other electrical equipment;
The power management method according to any one of appendices 11 to 19, wherein the characteristic section of the characteristic waveform is compared with a section corresponding to the characteristic waveform.

1 Current Measurement Unit 2 Current Waveform Data Acquisition Unit (Acquisition Means)
3 Recording unit 3a Storage device 4 Difference extraction unit (extraction means)
5 Model-specific current data management unit (management means)
6 Model estimation unit (estimation means)
7 Operating device counting unit 8 Output unit 8a Display device 10 Power management device 11 Feed line 14 Electric device

Claims (10)

Management means for storing the waveform of the operating current derived from the power supply circuit of the electrical device as a specific waveform;
Obtaining means for obtaining a waveform of a current supplied to the electrical device as a power supply waveform;
Extracting means for extracting a difference waveform between cycles of the power supply waveform;
An electric power management apparatus comprising: an estimation unit configured to estimate a model of the electric device whose operating state has changed by comparing the natural waveform with the differential waveform. The management means stores the characteristic waveform for one cycle of the operating current,
The acquisition means acquires the power supply waveform for a plurality of cycles of the supply current,
The power management apparatus according to claim 1, wherein the extraction unit extracts the difference waveform for at least one period from the power supply waveform for the plurality of periods. The estimating means is
When the similarity between the natural waveform and the differential waveform is high, it is estimated that the electrical device of the model corresponding to the natural waveform has been operated,
It is estimated that the electrical device of the model corresponding to the specific waveform has stopped operating when the similarity between the inverted specific waveform obtained by inverting the specific waveform with respect to the time axis and the differential waveform is high. The power management apparatus according to claim 1 or 2. The estimation means determines that the similarity is high when a sum of squares of a difference current corresponding to a waveform difference between the characteristic waveform and the difference waveform is less than a first threshold value. Item 4. The power management apparatus according to Item 3. 5. The method according to claim 1, wherein the extraction unit discards one of the power supply waveforms when the similarity between the pair of the power supply waveforms related to the extraction of the difference waveform is high. The power management apparatus according to item 1. The extraction means determines that the similarity between the power supply waveforms is high when the sum of squares of the difference currents corresponding to the waveform difference between the power supply waveforms is less than a second threshold, The power management apparatus according to claim 5. The extraction means excludes the power supply waveform including random fluctuations from the extraction target of the differential waveform based on the similarity between the power supply waveforms that are temporally changed. The power management apparatus according to claim 1. The extraction means for any of the power supply waveforms,
The sum of squares of the difference current corresponding to the waveform difference from the immediately preceding feeding waveform is calculated as the first sum of squares, and
While calculating the sum of squares of the difference current corresponding to the waveform difference from the power supply waveform immediately after as the second sum of squares,
The power management apparatus according to claim 7, wherein when the difference between the first sum of squares and the second sum of squares exceeds a third threshold, it is determined that the random fluctuation is included. The management means, for the electrical equipment having a plurality of operating modes, stores the unique waveform that differs for each operating mode,
The power management apparatus according to claim 1, wherein the estimation unit estimates the operation mode in addition to a model of the electrical device. The management means stores the characteristic waveform having a predetermined characteristic section which is at least a half wavelength or less and is distinguished from the operating current of the other electrical equipment,
The power management apparatus according to claim 1, wherein the estimation unit collates the characteristic section of the natural waveform with a section corresponding to the characteristic waveform.

Images