JP5010011B2 - Power consumption measuring system and power consumption measuring method - Google Patents

Description

  The present invention relates to a technique for individually measuring power consumption of a plurality of operating electric devices.

  As a technology to measure the power consumption of electric devices used in general households individually, the consumption of each electric device is based on the measurement of the lead-in line from the utility pole to the home instead of installing a measuring device for each electric device. A technique for estimating electric power has been proposed (for example, Patent Document 1).

  The system disclosed in Patent Document 1 includes a measurement sensor, data extraction means, and estimation means such as LMC. The data extraction means is the fundamental wave of the total load current from the measurement data VA, VB, IA, IB detected by the instrument transformer and the instrument current transformer installed near the service outlet of the power customer. In addition, data on harmonic current patterns and their phase difference patterns with respect to voltage are extracted. Then, this data is given as an input to the estimation means to estimate the individual power consumption of the electrical equipment.

International Publication No. 01/077696

  The technique disclosed in Patent Document 1 has a problem that the number of harmonic data combinations that need to be learned increases explosively when a plurality of electrical devices operate simultaneously. For example, when N electrical devices operate simultaneously, the number of combinations is 2 to the Nth power. Further, when each electric device has M harmonic patterns, combinations of harmonic data that need to be learned are 2 (N × M) powers. Specifically, for example, when 5 electric devices have 5 harmonic patterns, there are 2 25 (33,554,432) combinations. Although there are some fluctuations due to restrictions such as the fact that harmonic patterns of the same device do not appear at the same time, there is no doubt that the number of combinations of harmonic data that needs to be learned is enormous.

  It is extremely difficult to collect such a large number of harmonic data in advance, and a large-capacity memory for storing a large number of pattern data is required.

  The present invention has been made to solve the above-described conventional problems, and an object thereof is to provide a practical power consumption measurement system and a power consumption measurement method capable of accurately estimating the power consumption of each electrical device. And

In order to achieve the above object, a power consumption measurement system according to the present invention includes:
Based on the total load current and voltage measured at predetermined locations on the service area, the current waveform data averaged over one cycle of the commercial frequency of the total load current is extracted, and the current value is extracted from the averaged current waveform data. A data extraction means for extracting convex point information on a convex point indicating a point at which the change of the point changes from an increase to a decrease or a point at which the change changes from a decrease to an increase;
Based on the convex point information extracted by the data extraction means and the estimated model, an estimation model that associates the type of electrical equipment, the convex point information, and power consumption is stored in advance. And estimation means for individually estimating the power consumption of the electrical equipment.

  According to the present invention, it is possible to accurately estimate the power consumption of each electrical device using an estimation model that is easy to construct.

It is a block diagram which shows the structure of the power consumption measuring system which concerns on one Embodiment of this invention. It is a figure for demonstrating the installation aspect of the power consumption measuring system of FIG. It is a figure for demonstrating the measuring instrument transformer and measuring instrument current transformer of FIG. It is a block diagram which shows the structure of the measurement part of FIG. It is a block diagram which shows the structure of the data extraction part of FIG. (A) is a graph which shows the example of a waveform of the current data obtained by the A / D conversion part of FIG. 5, (b) shows the example of a waveform of the current data preserve | saved at the waveform storage memory of FIG. It is a graph. It is a figure for demonstrating convex point information. It is a flowchart which shows the procedure of an estimation model production | generation. It is a figure which shows the structure of a registration circuit. It is a graph which shows an example of the current waveform of an IH cooking heater. It is a graph (the 1) which shows an example of the current waveform of an air conditioner. It is a graph (the 2) which shows an example of the current waveform of an air conditioner.

  Hereinafter, a power consumption measurement system according to an embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of a power consumption measurement system 1 according to the present embodiment. The power consumption measurement system 1 includes a measurement unit 10, an instrument transformer 20, and an instrument current transformer 30. As shown in FIG. 2, the power consumption measurement system 1 is installed in the vicinity of a distribution board 2 in a power user's house (demand area), and measures individual power consumption of a plurality of electric devices 3 used in the house. To do.

  The distribution board 2 is connected to an electric power system such as an electric power company through electric wires (not shown) installed on the lead-in wire 4 and the utility pole 5. The distribution board 2 is connected to a power supply line 6 for supplying power to the electrical equipment 3. In the present embodiment, an inverter device such as an air conditioner (air conditioner) 3a, a microwave oven 3b, an IH (Induction Heating) cooking heater 3c, and a washing machine 3d is connected to the power supply line 6 and used as the electric device 3. Shall.

  As shown in FIG. 3, the instrument transformer 20 is composed of an A-phase instrument transformer 201a and a B-phase instrument transformer 201b, and the instrument current transformer 30 is an A-phase transformer. Instrument current transformer 301a and B phase instrument current transformer 301b.

  The instrument transformer 201a has a primary side connected between the A phase 4a and the neutral wire 4n, and outputs a voltage VA similar to the voltage of the A phase 4a from the secondary side. The secondary side of the instrument transformer 201a is connected to the measurement unit 10 via a connection line 202a such as a coaxial cable. The instrument transformer 201b has a primary side connected between the B phase 4b and the neutral wire 4n, and outputs a voltage VB similar to the voltage of the B phase 4b from the secondary side. The secondary side of the instrument transformer 201b is connected to the measurement unit 10 via a connection line 202b such as a coaxial cable.

  The instrument current transformer 301a measures the current flowing through the A phase 4a on the primary side and outputs a current IA similar to the A phase current from the secondary side. The secondary side of the instrument current transformer 301a is connected to the measurement unit 10 via a connection line 302a such as a coaxial cable. The instrument current transformer 301b measures the current flowing in the B phase 4b on the primary side and outputs a current IB similar to the B phase current from the secondary side. The secondary side of the instrument current transformer 301b is connected to the data measuring unit 10 through a connection line 302b such as a coaxial cable. In the present embodiment, an instrument current transformer having a through-type or clamp-type structure is adopted as the instrument current transformers 301a and 301b.

  As shown in FIG. 4, the measurement unit 10 includes a data extraction unit 11, an estimation unit 12, and a communication unit 13. The communication unit 13 communicates with a registration circuit, which will be described later, and performs processing for receiving teacher data (details will be described later) transmitted from the registration circuit. The data extraction unit 11 extracts data for one cycle of the commercial frequency of the total load current based on the voltage and current output from the instrument transformer 20 and the instrument current transformer 30. In this embodiment, the commercial frequency is 50 Hz. Therefore, one period is 20 milliseconds.

  More specifically, as shown in FIG. 5, the data extraction unit 11 includes a low-pass filter 111 that cuts low frequency components, an analog / digital (A / D) conversion unit 112, a waveform storage memory 113, and a convex inspection. And an exit part 114.

  The low pass filter 111 amplifies the current signal after cutting the low frequency component of the current signal. Thereby, the measurement accuracy by the A / D conversion unit 112 is improved. Specifically, the low-pass filter 111 reduces a strong low-frequency signal that hinders detection of a convex point described later, amplifies the extracted high-frequency signal, and measures the A / D conversion unit 112. Amplifies to full range. Thereby, the quantization error of sampling can be reduced.

  The A / D converter 112 samples the currents IA and IB at 20 KHz based on the voltages VA and VB, respectively, converts them into digital data, and then records them in the waveform storage memory 113. Here, the current waveform obtained by the A / D converter 112 is a waveform that repeats every section (20 milliseconds in the case of 50 Hz) as shown in FIG. In this case, generally, noise or the like has a characteristic that is not repeated for each section. The sampling rate of the A / D conversion unit 112 is not necessarily 20 KHz, and equivalent performance can be obtained as long as it is several kilohertz or more.

  Specifically, the A / D converter 112 superimposes data (current waveform data of each of the A phase and B phase) in one section (one commercial frequency cycle) obtained by the conversion in time series. Record in the waveform storage memory 113. For example, each current waveform data in the sections 401 to 403 as shown in FIG. 6A is sequentially accumulated, and as shown in FIG. 6B, the accumulated current 404I is stored in the waveform accumulation memory 113. The Here, the integrated current 404I is data obtained by adding the currents (currents 401I, 402I, 403I) of each section in time series.

  In addition, the A / D converter 112 also records the number of times of current waveform data for one section (that is, the number of integrations) in the waveform storage memory 113.

  The convex point detection unit 114 reads the accumulated current waveform data of each of the A phase and the B phase recorded in the waveform storage memory 113 and divides the accumulated current waveform data by the respective number of times of accumulation, thereby obtaining an average of each of the A phase and the B phase. Current data is calculated, and data relating to each convex point is detected.

  The convex point means a point at which the change in the time axis direction of the current value changes from increase to decrease, or a point at which the change starts from decrease to increase, and appears as a convex protrusion on the current value graph. Hereinafter, the difference between the current value at which the convex point occurs and the moving average value is referred to as a convex height, and the distance between the intersections of the current value and the moving average value across the convex point is referred to as a convex width. In FIG. 7, the convex point 60 indicates a point at which the time series change of the current has changed from increase to decrease, or a point at which the current has changed from decrease to increase. The position of the convex point 60 is indicated by a position on the time axis. The convex point 60 has two feature amounts, a convex height 61 and a convex width 62, which are determined by the degree of change in increase and decrease.

  In the example of FIG. 7, the convex height 61 is the absolute value of the difference between the current value 6 a and the moving average value 6 b at the convex point 60 with respect to the current value 6 a and the moving average value 6 b of the current value. The convex width 62 is an absolute value of a difference in the time axis direction between two intersections (intersection 63 and intersection 64) of the current value 6a and the moving average value 6b.

  Here, the moving average value 6b is an average value of N samples before and after the current value 6a in the time axis direction. In addition, about the value of N, you may set a different value for every electric equipment 3 of estimation object, and it is good also as a fixed value. Moreover, it is good also as a weighted average which weights a sample instead of making it a uniform average value.

  The estimation unit 12 executes an estimation algorithm using an estimation model generated based on pre-registered teacher data, and inputs the position, convex height, and convex width of each convex point detected by the convex point detection unit 114 As a result, the individual power consumption of the electrical device 3 in the operating state is estimated.

  The position of the convex point, the convex height, and the convex width are characterized in that observation is easy even in the total load current obtained by joining all the load currents in a situation where a plurality of electrical devices 3 are operating. For this reason, it is possible to easily determine whether or not the target electrical device 3 is operating.

  When the power consumption of the inverter device changes (for example, when the set temperature is changed in the air conditioner 3a), the position of the convex point does not change, and the convex height and the convex width are proportional to the power consumption. Thus, individual power consumption can be estimated from the convex width and convex height. Since it is rare that the positions of the convex points collide with each other in the plurality of electric devices 3, even if the plurality of electric devices 3 are operating simultaneously, the power consumption of the individual electric devices 3 can be estimated. The teacher data to be registered in the estimation unit 12 can be generated by measuring only the load current of the electric device 3 alone without considering the combination of the electric devices 3. Therefore, the labor required for registering teacher data can be greatly reduced.

  The estimation unit 12 holds an estimation model generated based on teacher data. And when the data regarding the convex point, convex width, and convex height of the average current waveform of the actually measured total load current are inputted, the individual power consumption of the electric equipment 3 is calculated using the above estimation model. presume.

  A procedure for generating the estimation model will be described with reference to FIG.

  First, on the basis of the load current obtained by individually measuring the electrical device 3 to be estimated by a registration circuit described later, convex point data in the electrical device 3 is acquired (step S101). Here, the registration circuit acquires convex point data for each operation state of the electrical device 3.

  The data of the position of the convex point, the convex width, and the convex height are preferably normalized in order to equalize the weight of the information because the range of each value is different. For example, with respect to the position of the convex point, the registration circuit sets the maximum time (20 milliseconds) for the phase from the voltage to [0-1] and the convex width similarly to [0-1]. For normalization, normalization is performed to convert the maximum input (± 10 A) of the A / D conversion unit 112 to [−1 to +1] as a maximum value (step S <b> 102).

  Next, the registration circuit associates the teacher data composed of the normalized data and the power consumption of the operating electric device 3 as the answer at that time with each type of the electric device 3 in association with the type of the electric device 3. To the communication unit 13 of the measurement unit 10. Thereby, teacher data for each operation state of the electrical device 3 is registered in the estimation unit 12 of the measurement unit 10 (step S103).

  The estimation unit 12 estimates the power consumption curve of the electric device 3 with respect to the convex width and convex height from the teacher data for each operation state by multiple regression analysis (step S104). Specifically, the parameter a and the parameter b in Equation 1 below are determined so that the error is minimized. The least square method is used as the error calculation method. And the estimated power consumption in each convex point for every operation state is calculated | required by Formula 1 (step S105).

ｉ＝１〜ｍ、
α：ずれの許容範囲を示す定数、
ｄ：基準凸点位置とのずれ

i = 1 to m,
α: constant indicating the allowable range of deviation,
d: Deviation from the reference convex point position

  Next, as shown in Equation 2 below, the power consumption estimated from each convex point is integrated for each operation state, and an average value is calculated (step S106). In addition, from the result of Formula 1, those with estimated power consumption of 0 are excluded from the average calculation of Formula 2. In addition, the estimated power consumption of a convex point with a low contribution is also excluded from the average calculation of Equation 2.

  In the present embodiment, from the result of Equation 1 above, when the number of convex points whose estimated power consumption is 0 in any of the operating states exceeds 50% of the total, The power consumption in the operating state cannot be estimated. What percentage is used as a threshold is arbitrary.

  As described above, an estimation model for estimating the power consumption in each operation state of one electrical device 3 is generated. And the estimation model for estimating the power consumption in each operation state of each electric equipment 3 is acquirable by performing the process mentioned above about all the electric equipment 3 made into an estimation object.

  As described above, by calculating the power consumption estimation formula based on the multiple regression equation, the power consumption can be estimated even when the electrical device 3 is in an unknown operating state. That is, for example, in the case of the air conditioner 3a, since the power consumption continuously changes, the combinations of the power consumption and the respective convex point positions, convex heights, and convex widths are generated infinitely. Although it is practically impossible to register all of these, it is necessary to register all combinations by calculating a linear prediction formula based on multiple regression analysis as in the power consumption measurement system 1 of the present embodiment. The registration work becomes very easy.

  Next, a registration circuit for registering teacher data in the estimation unit 12 will be described.

  As shown in FIG. 9, the registration circuit includes a sine wave power supply device (50 Hz, 100 V) 51, a switch 52, a current transformer 53 for measuring the total load current, and a voltage for measuring the voltage. An instrument transformer 54, a waveform converter 55, and a communication unit 56 are included. In addition, an air conditioner 3a, a microwave oven 3b, an IH cooking heater 3c, and a washing machine 3d are selected as electric devices 3 to be registered.

  In the air conditioner 3a, the microwave oven 3b, the IH cooking heater 3c, and the washing machine 3d, the power consumption changes very finely by the inverter. Therefore, by changing the operating state of these electrical devices 3 (for example, in the case of the air conditioner 3a, the indoor set temperature and the set wind speed are changed), teacher data in various usage situations (operating states) of these electrical devices 3 Can be supplied to the estimation unit 12.

  Since the communication unit 56 communicates with the communication unit 13 of the measurement unit 10, the registration circuit can supply teacher data while the measurement unit 10 is installed on the distribution board 2.

  Next, the convex points, the convex heights, and the convex widths of the operating state of the electrical equipment 3 and the average current waveform of the load current at that time will be described in detail. FIG. 10 shows an example of an average current waveform when the IH cooking heater 3c is used with a heating power of medium heat (1600 W). In FIG. 10, one of the convex points exists at a position of 0.0125 as a normalized position 250 microseconds after the point where the voltage becomes zero. At this time, the convex height is 0.046, and the convex width is 0.0175. Similarly, other convex points exist at positions indicated by 0.035, 0.0525, and 0.0725, respectively.

  For example, when the air conditioner 3a is operating in the heating mode (for example, air volume: medium) (800 W), an average current waveform and a convex point as shown in FIG. 11 are obtained. The positions of the convex points shown in FIG. 11 (position 0.1075, position 0.1325) are significantly different from the case of the IH cooking heater 3c shown in FIG. As described above, the different electrical devices 3 are accompanied by a specific average current waveform and a convex point corresponding to the operation state. These convex point information can be easily extracted even when the currents of other electrical devices not to be estimated are mixed. For example, even when the load current of the IH cooking heater 3c and the load current of the illuminator are combined, the convex point information of the IH cooking heater 3c can be easily identified.

  As described above, the operation state of each electric device 3 can be easily detected by collating the position of the convex point, the convex width, and the convex height in various operation states of the electric device 3. In addition, even when a load current corresponding to another electrical device is mixed, the operation state can be detected.

  FIG. 12 shows an example of an average current waveform when the air conditioner 3a is operating at a load of 1000W and when operating at a load of 1400W. FIG. 12 shows a change in load current for one voltage cycle. The left end of the horizontal axis in FIG. 12 is a point where the voltage is 0 V (phase 0), and the right end is a point where the phase is 360 degrees. It is. In the present embodiment, since the voltage frequency is 50 Hz, the right end indicates the load current value at the time when 20 milliseconds have elapsed from the left end.

  In FIG. 12, when comparing the average current waveform when operating at 1400 W and when operating at 1000 W, the convex point position does not change in either case, and the convex height and convex width. It can be seen that increases in proportion to the power of the load. However, not all convex points increase with the same proportionality constant, and the proportionality constants differ for each convex point. Therefore, as described above, for each convex point, it is necessary to obtain a regression equation for estimating power consumption from the convex width and convex height by multiple regression analysis.

  As described above, the estimation unit 12 is operating according to an estimation model that receives data on the position, height, and width of the convex point of the average current of the total load current and outputs the power consumption of the electrical device 3 as an output. The power consumption of the electrical equipment 3 is estimated individually. Therefore, even if current waveform data having a height or width of a convex point that was not detected at the time of registration is detected, the power consumption of the electric device 3 can be estimated.

  When the positions of the convex points do not match, the estimation unit 12 treats the electrical device as something other than the registration target (that is, the estimation target) (unknown electrical device), and estimates the power consumption. Absent. However, even if one or more unknown electrical devices are operating, the power consumption of the electrical device 3 to be registered can be estimated.

  As described above, according to the power consumption measurement system 1 of the present embodiment, the individual power consumption of the electrical equipment 3 actually used in the home of the power user is installed for each electrical equipment 3. It can measure without.

  For example, the estimated power consumption data of the electrical device 3 may be transmitted from the measurement unit 10 to a device including the above-described registration circuit, and displayed on the device to be presented to the user. Moreover, you may make it transmit to the server which service providers, such as an electric power provider, operate via a predetermined | prescribed communication line. In this manner, various information services are provided from the service provider to the power user, and information on the power user side is also collected by the service provider through the network and is effectively used to construct a new service. be able to.

  For example, one of the important information for electric power companies is information on the configuration and actual usage of electric devices owned by each electric power user. In particular, in commercial facilities such as buildings, energy saving can be promoted by obtaining a breakdown of the actual usage of electricity. The power consumption measurement system 1 of the present embodiment can be said to be one of powerful systems that can meet such needs.

  In the present embodiment, the estimation of the individual power consumption of power devices in a general household has been described. In this case, for example, by monitoring and recording the operating state and power consumption of a specific electrical device in the building, what amount of power consumption of the entire building is brought about by the specific electrical device at a later date. It can be analyzed whether it was made.

  In addition, this invention is not limited to the said embodiment, Of course, the various change in the range which does not deviate from the summary of this invention is possible.

  For example, in the above embodiment, the estimation unit 12 uses an estimation algorithm such as multiple regression analysis for power consumption estimation. However, the estimation unit 12 is not limited to this. For example, an estimation formula using a neural network or a genetic algorithm is used. May be used.

  As an estimation method using a neural network, it is preferable to use a hierarchical neural network having a feature that an estimation model can be established through sequential learning. With this estimation method, the prediction system of the estimation formula can be gradually improved after starting to use.

  In addition, when a genetic algorithm is used, the prediction formula itself can be flexibly evolved, and even in a model in which the linearity of the prediction formula implicitly assumed in the above-described formulas 1 and 2 is not satisfied, estimation is performed. A highly accurate prediction formula can be obtained.

  Further, an estimation model of a specific electrical device (for example, provided by the manufacturer of the electrical device) may be additionally registered in the estimation unit 12 using a telephone line, an optical fiber dedicated line, or the like. In this way, it is not necessary for each power user to register teacher data and the convenience is improved.

  The present invention can be suitably used not only in general households but also in buildings, factories, and the like as a system for measuring individual power consumption of electrical equipment.

DESCRIPTION OF SYMBOLS 1 Power consumption measuring system 10 Measurement part 11 Data extraction part 111 Low pass filter 112 A / D conversion part 113 Waveform accumulation memory 114 Convex point detection part 12 Estimation part 13 Communication part 20 Instrument transformer 201a Instrument transformer (for A phase) )
201b Instrument transformer (for phase B)
202a, 202b Connection line 30 Current transformer for instrument 301a Current transformer for instrument (for phase A)
301b Current transformer for instrument (for B phase)
302a, 302b Connection line 2 Distribution board 3 Electrical equipment 4 Lead-in line 5 Utility pole 6 Power supply line 51 Sine wave power supply device 52 Switch 53 Instrument current transformer 54 Instrument transformer 55 Waveform conversion part 56 Communication part

Claims (4)

Based on the total load current and voltage measured at predetermined locations on the service area, the current waveform data averaged over one cycle of the commercial frequency of the total load current is extracted, and the current value is extracted from the averaged current waveform data. A data extraction means for extracting convex point information on a convex point indicating a point at which the change of the point changes from an increase to a decrease or a point at which the change changes from a decrease to an increase;
Based on the convex point information extracted by the data extraction means and the estimated model, an estimation model that associates the type of electrical equipment, the convex point information, and power consumption is stored in advance. An estimation means for individually estimating the power consumption of the electrical equipment in
A power consumption measurement system characterized by that. The convex point information includes information about the position, width and height of the convex point.
The power consumption measuring system according to claim 1. The estimation model is represented as a linear model by a multiple regression equation with the convex point information as an explanatory variable and the electric power consumption of each electric device as an objective variable.
The power consumption measuring system according to claim 1 or 2. Based on the total load current and voltage measured at predetermined locations on the service area, the current waveform data averaged over one cycle of the commercial frequency of the total load current is extracted, and the current value is extracted from the averaged current waveform data. A data extraction step of extracting convex point information relating to a convex point indicating a point at which the change of the point changes from increasing to decreasing, or a point at which decreasing changes to increasing;
Based on the estimation model that associates the type of electric device, the convex point information, and the power consumption, and the convex point information extracted in the data extraction step, the power consumption of the electric device in operation is calculated. An estimation step for individually estimating,
A method for measuring power consumption.

Images