KR102079916B1 - System and method for cost-effective notification of home appliance usage considering billing rate and power consumption pattern - Google Patents

Abstract

According to an exemplary embodiment, a usage notification method performed by a usage notification system includes analyzing and predicting a power consumption pattern through machine learning about power consumption; And providing estimated power rate information within a charging period linked to a progressive rate based on the predicted power consumption pattern.

Description

SYSTEM AND METHOD FOR COST-EFFECTIVE NOTIFICATION OF HOME APPLIANCE USAGE CONSIDERING BILLING RATE AND POWER CONSUMPTION PATTERN}

The following description relates to a system and method for analyzing a power consumption pattern for each time zone and notifying a user to use an efficient home appliance in consideration of a progressive rate.

In the use of home appliances, power is wasted due to the driving of devices not recognized. This waste of power accumulates and accumulates, leading to an increase in the total amount of power consumed per month, especially if unintentional operation is being made for devices that require high power, such as summer air conditioners or winter electric heaters. Increases. In addition, this unrecognized and wasting power often causes electricity bills to be paid at the price of an unwanted electricity progressive tax.

Although the monitoring system can identify the amount of power consumed and the user's judgment, the overall power consumption can be planned.However, if each device does not know the contribution to the power consumption, it will be possible to develop an efficient operation plan for each device in detail. Difficult to build

The Energy Management System (EMS) is an enterprise-wide system for achieving energy efficiency by efficiently managing and saving energy, and is emerging as an energy resource. Among them, monitoring technology, which is an important element of energy management system, and prediction technology based on energy usage pattern are being used in various ways. In this regard, Korean Patent Laid-Open Publication No. 10-20140132523 relates to a power load monitoring apparatus and method, and discloses a power consumption monitoring technology for detecting a load change and identifying a load state through power consumption information obtained through a sensor. International Patent Publication No. 10-1631781 discloses a system for predicting future demand power through power consumption data.

By using machine learning, it analyzes the pattern of power used at each time for the electronic device used in a specific environment and predicts the amount of power used in the billing period to provide notification information for the efficient use of the electronic device considering the electric leakage error rate. We propose a method and system.

The usage notification method performed by the usage notification system may include analyzing and predicting a power consumption pattern through machine learning about power consumption; And providing estimated power rate information within a charging period linked to a progressive rate based on the analyzed and predicted power consumption pattern.

Providing the estimated power rate information in the charging period linked to the progressive rate based on the analyzed and predicted power consumption pattern, by adjusting the consumption of each device for the analysis and predicted power consumption pattern through a simulation And calculating the final billing details in advance.

Based on the analysis and the predicted power consumption pattern, providing the estimated power rate information in the charging period linked to the progressive rate, the power consumption pattern for the remaining charging period of the charging period based on the predicted power consumption pattern. Proposing steps may include.

Analyzing and predicting a power consumption pattern through the machine learning about the power consumption, classify the power consumption information for each electronic device through the machine learning from the total power consumption for the electronic devices used in a specific environment It may include the step.

Analyzing and predicting a power consumption pattern through the machine learning about the power consumption, the consumption pattern of each time period from the power consumption information for each electronic device or the power consumption pattern from the power consumption information for each electronic device It may include obtaining any one or more.

Providing the estimated power rate information within the charging period linked to the progressive rate based on the analyzed and predicted power consumption pattern may include the proposed power consumption pattern, the amount of power accumulated up to now from each electronic device, and within the charging period. The method may include notifying the total cumulative amount of power expected, the amount of power accumulated so far from electronic devices existing in a specific environment, or notification information including the estimated power rate information.

The usage notification system includes a prediction and analysis unit for analyzing and predicting a power consumption pattern through machine learning about power consumption; And a provision unit for providing estimated power rate information within a charging period linked to a progressive rate based on the analyzed and predicted power consumption pattern.

Instead of simply planning the operation of power consumption with information on the amount of power consumed by the electronic device, it effectively monitors the proportion of each device's power consumption and whether unintended operation of the device is effective. Can stand. It can also reduce wasted power and manage power usage for progressive trends.

1 is a block diagram illustrating a configuration of a usage notification system according to an exemplary embodiment.
2 is a flowchart illustrating a method of learning a pattern for power consumption based on machine learning in a usage notification system according to an exemplary embodiment.
3 is a flowchart illustrating a method of performing a simulation in a usage notification system, according to an exemplary embodiment.
4 is a flowchart illustrating a method for managing power consumption efficiency considering a progressive rate and a power consumption pattern according to time zones in a usage notification system according to an exemplary embodiment.
5 illustrates an example of determining power consumption pattern information for each time zone in a usage notification system according to an exemplary embodiment.
6 illustrates an example of determining power consumption pattern information in a usage notification system according to an exemplary embodiment.
7 and 8 are examples of simulation of a usage notification system according to an embodiment.
9 is an example for describing a method of performing preprocessing and machine learning in a usage notification system according to an exemplary embodiment.

Hereinafter, exemplary embodiments will be described in detail with reference to the accompanying drawings.

In the following examples, preprocessing and machine learning (deep learning) techniques are used to monitor electronic devices used in specific places (eg, homes) in real time and to enable more efficient power consumption management than simple monitoring. An efficiency management system and method based on power consumption information will be described.

1 is a block diagram illustrating a configuration of a usage notification system according to an embodiment, and FIG. 4 is a diagram for managing power consumption efficiency considering a progressive rate and a power consumption pattern according to a time slot in a usage notification system according to an embodiment. It is a flow chart.

The usage notification system 100 may include an analysis and prediction unit 110 and a provider 120. These components may be representations of different functions performed by a processor in accordance with control instructions provided by program code stored in the usage notification system 100. In this case, the components may be implemented to execute instructions according to code of an operating system included in the memory and code of at least one program.

The processor of the usage notification system 100 may load a program code stored in a file of a program for a method for managing power consumption efficiency in consideration of a progressive rate and a time-phase power consumption pattern in a memory. For example, when a program is executed in the usage notification system 100, the processor may control the usage notification system to load program code from a file of the program into a memory under control of an operating system. In this case, each of the processor and the analysis and prediction unit 110 and the provider 120 included in the processor executes instructions of a corresponding part of the program code loaded in the memory to execute the subsequent steps 410 to 445. It may be different functional representations of the processor.

The analysis and prediction unit 110 may analyze and predict the power consumption pattern using machine learning. The analysis and prediction unit 110 may acquire the total power consumption of the specific environment (410). The power consumption of a particular environment, for example the entire electronic device present in the home, can be measured integrally. For example, the smart meter may be provided with total power consumption for all electronic devices at regular intervals. The analysis and prediction unit 110 may perform machine learning on power consumption (420). For example, the RNN model may be applied to learn about time-specific consumption patterns and power consumption patterns of each electronic device. In this case, after a sufficient period of time, the power consumption pattern for each electronic device may be generated through the discriminator 1, the discriminator 2, and the predictor through the learned models. By applying Non-Intrusive Load Monitoring (NILM) to the total power consumption received from the smart meter, power consumption for each electronic device can be derived. NILM classifies and monitors power data of several independent nodes from a single power data mixed with various power data information, and can be implemented using various types of machine learning techniques such as HMM and SVM. .

The analyzing and predicting unit 110 may obtain power consumption for each electronic device by pattern discriminator 1 for each time slot. The pattern discriminator 1 determines whether the power consumption of each electronic device matches the learned time-based consumption pattern. For example, the pattern discriminator 1 may determine whether power consumption is generated within a specific time zone range, and pattern the power consumption within a specific time zone range to derive time-specific consumption patterns. If the providing unit 120 occurs outside a specific time zone, the providing unit 120 may transmit information on the corresponding device to the user. For example, when there is a time when an electronic device is generally used, and suddenly it is used in another time zone, this information can be delivered to the user in real time. Specifically, the power supply of the electronic device may be checked as the electronic device is used when the electronic device is not used. In addition, the analysis and prediction unit 110 may acquire the power consumption pattern by the pattern discriminator 2 as the consumption pattern for each time zone is generated by the pattern discriminator 1 (442). In this case, a power consumption pattern for power consumption for each electronic device may be derived. For example, the pattern discriminator 2 may determine whether the power consumption pattern for the electronic device is normal power (443). When the power consumption pattern is consumed at normal power, the provider 120 may terminate the process or deliver information normally consumed to the user. In addition, when the power consumption pattern is not consumed at normal power (for example, a preset power range), the provider 120 may transmit information of an electronic device abnormally consumed to the user. In this case, information including identification information of an abnormally consumed electronic device, a power consumption amount of the abnormally consumed electronic device, and the like may be notified to the user.

Meanwhile, the analysis and prediction unit 110 may acquire the total cumulative power consumption predicted within the billing period by the pattern predictor 430 as the machine learns the power consumption of the entire specific environment (431). In this case, the amount of cumulative power (use) predicted for each electronic device may be provided together during the charging period. In other words, the pattern predictor may predict the amount of power each household appliance will consume for the remaining period of the charging period. If the estimated cumulative power amount is larger than the progressive rate reference, the provider 120 may transmit the information and the recommended solution about the electronic device that needs to adjust the consumption pattern (433). Through the predicted power consumption of each electronic device, the total amount of power to be consumed for the remaining period of the charging period can be derived, and the total cumulative power consumption within the charging period can be predicted by adding up the total accumulated power in real time. In addition, when the total cumulative consumption in the estimated billing period is compared with the progressive tax reference table, and it is determined that additional progressive tax will occur, the adjustment for the electronic device which occupies a high proportion of the amount of electricity exceeded from the progressive tax standard and the total power consumption or consumes high power The notification information can be delivered to the user. For example, the proposed power consumption pattern from the present to the remaining billing period, the amount of power accumulated so far from each electronic device, the total accumulated power amount expected within the billing period, and the cumulative power accumulated from the electronic devices present in the specific environment to the present. Notification information including the amount of electricity or the estimated power rate information may be notified.

Through machine learning, the user can be notified of information on the currently running electronic device (eg, home appliances) and power consumption in real time. When it is determined that the drive is out of the power consumption time pattern of the device for each electronic device and it is determined that an unintended drive has occurred, the user may be notified to prevent the unintentional drive of the electronic device from continuing. . The user may take action on the electronic device to reduce power consumption unintentionally wasted. When driving within the power consumption time pattern, but having a different power consumption pattern than usual, it is determined that a singularity has occurred in the corresponding electronic device, and the user is notified of the notification information to enable preliminary measures.

In addition, it is possible to determine the total cumulative power consumption in real time and at the same time predict the future power consumption to determine whether the total monthly power consumption meets the progressive tax criteria. If it is predicted that further progress will be applied, it can inform the user that adjustments need to be made for devices that account for a high proportion of the total power consumption or consume high power.

2 is a flowchart illustrating a method of learning a pattern for power consumption based on machine learning in a usage notification system according to an exemplary embodiment.

The usage notification system may acquire power consumption information of each electronic device in the home (210). In this case, the usage notification system acquires total power consumption information for the electronic appliances in the home. The usage notification system may learn a power consumption pattern for each electronic device through machine learning (220). For example, one or more of time-specific consumption patterns or power consumption patterns may be acquired as the power consumption pattern.

At this time, the usage notification system performs a preprocessing process, and based on the RNN (Multi-layer LSTM based RNN) model among the deep learning models, it determines the abnormality of the classification technology and power usage pattern of the NILM and the power consumption (use) amount. It can be predicted. Referring to FIG. 9, before learning data using the machine learning (deep learning) 900, the feature of the data may be extracted using an audio signal processing technique for efficient learning of the data. In this regard, a technology that considers frequency and time components of data such as spectrogram, MFCC, and Mel-spectrogram can be applied.

The general RNN model derives the result by considering not only current time data information but also data of previous time. The derived result has a desired type of result. In the embodiment, the result of the classification result of NILM, the presence or absence of a power usage pattern abnormality, and the predicted power consumption value are the result values. In general, although the learning form of the RNN model is various, it is possible to learn the results by selecting the learning form according to the purpose of the invention.

The classification technique of NILM, that is, Mutil label classifiation and separation using RNN can be used for disaggregation of power data. At this time, Mutil label classifiation is learned in many to one manners, and separation is learned in many to many manners based on the results of Mutil label classifiation.

The usage notification system may determine whether the accuracy of the generated pattern for the power data is above or above the threshold (230). If the pattern generated for the power data is less than / less than the threshold value, the usage notification system may acquire or measure power consumption information again to learn the pattern. If the pattern generated for the power data exceeds the threshold, the usage notification system may determine whether the pattern is abnormal (240).

In order to determine whether there is an abnormality in the power usage pattern, the RNN model can learn in many to one ways with two results including the presence or absence of a pattern abnormality for the power data. Here, the type of abnormality may be performed with respect to the time zone pattern of FIG. 5 and the pattern of power consumption of FIG. 6. For example, referring to FIG. 5, a time-based consumption pattern of an electronic device may be derived as learning is performed based on power consumption information. The abnormal pattern information may be determined by determining power consumption information of the same electronic device based on the time-specific consumption pattern information. Referring to FIG. 6, a power consumption pattern of an electronic device may be derived as learning is performed based on power consumption information. The abnormal pattern information may be determined by determining power consumption information of the same electronic device based on the power consumption pattern of the electronic device. Power usage prediction can predict the amount of power to be used in the future by learning the previous power usage pattern in many to one manners.

According to the embodiment, in order to solve the limitation of NILM based on the existing classification technology such as HMM, deep learning technology is applied, and data feature extraction through signal processing is performed as a preprocessing process. The RNN model is applied as a classification technique to enable higher performance NILM.

3 is a flowchart illustrating a method of performing a simulation in a usage notification system, according to an exemplary embodiment.

A method of performing a simulation provided by a usage notification system will be described with reference to FIGS. 7 to 8.

The usage notification system may perform machine learning on the total power consumption 310 in the home (310). In general, the total power consumption resulting from at least one or more electronic devices present in the home can be obtained. By performing machine learning on the total power consumption, a power consumption pattern for each electronic device can be predicted (330).

The usage notification system may provide a simulation for predicting the total power consumption during the charging period (340). Referring to FIG. 7, the usage notification system may provide a total cumulative power amount in which power consumption and power consumption for each electronic device are accumulated. In addition, a predicted cumulative amount of power exceeded during the charging period may be provided for the power consumption of each electronic device.

A device may be selected for adjusting the power consumption from the user. For example, the usage notification system may provide a user interface for performing a simulation. Specifically, the usage notification system may provide a user interface for inputting condition information for simulation by the user, and an operation by the user may be performed through the provided user interface. For example, a device for adjusting power consumption may be selected or input through a user interface. Referring to FIG. 8, the user may select some of the electronic devices with high usage amount as the total accumulated power amount is determined to be exceeded. In this case, the usage time of the selected electronic device may be adjusted.

Equation 1 for predicting and simulating the total amount of power used during the charging period may be expressed as follows.

), N은 사용자가 조정할 기기의 수,

는 사용자가 조정할 기기 n의 전력 사용량 조정 비율(

),

는 시뮬레이션 전, 과금 기간 내 예측된 전자 기기 n의 전력 소비량,

는 시뮬레이션 후, 과금 기간 내 예측된 전자 기기n의 전력 소비량,

은 전자 기기 n의 현재까지 누적된 전력 소비량,

은 시뮬레이션 결과를 의미할 수 있다. 이와 같이, 수학식 1에 기반한 시뮬레이션이 수행됨으로써 과금 기간 동안의 총 전력 소비량을 예측할 수 있다. Where T is the last time of the billing period and t ₀ is the simulation start time (

), N is the number of devices you want to adjust,

Is the percentage of power usage adjustment for device n

),

Is the estimated power consumption of electronic device n during the charging period,

After simulation, the estimated power consumption of electronic device n in the charging period,

Is the cumulative power consumption of electronic device n,

May mean a simulation result. As such, the simulation based on Equation 1 is performed to predict the total power consumption during the charging period.

Through Equation 2, the final billing amount calculation simulation can be performed within the billing period.

는 s 단계 구간 내에서 사용한 전력량(

),

는 전기 요금표 s 단계의 구간 기준 요금(

),

는 전기 요금표 s 단계의 구갠 최대 제한 전력량(

),

는 기본 요금, R은 최종 과금 금액을 의미할 수 있다. 이와 같이, 수학식 2에 기반한 시뮬레이션을 통하여 과금 기간 내에 최종 과금 금액을 미리 계산할 수 있다. Where s is the electricity rate step, S is the electricity rate step corresponding to the total amount of power used,

Is the amount of power used within the s phase

),

Is the interval-based rate for step s of the electricity rate table (

),

Is the maximum limit of the amount of power consumed

),

Is the basic charge, and R may mean the final billing amount. As such, the final billing amount can be calculated in advance within the billing period through a simulation based on Equation 2.

The usage notification system may obtain the estimated cumulative power (use) amount through the simulation (350). In other words, the usage notification system may predict the total cumulative power (use) amount during the charging period as the simulation is performed, and may calculate the billing amount for the total cumulative (use) amount in advance. For example, simulation can reduce or increase the use of a particular device in order to reduce the total cumulative amount. By manipulating the consumption for this particular device, you can reduce the total cumulative power and cope with it to avoid the progressive tax associated with the total cumulative power. In addition, it is possible to estimate the power consumption to be used for the remaining billing period until the end of the billing period based on the cumulative amount of power so far based on the billing start date during the billing period.

The apparatus described above may be implemented as a hardware component, a software component, and / or a combination of hardware components and software components. For example, the devices and components described in the embodiments are, for example, processors, controllers, arithmetic logic units (ALUs), digital signal processors, microcomputers, field programmable gate arrays (FPGAs). May be implemented using one or more general purpose or special purpose computers, such as a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to the execution of the software. For convenience of explanation, one processing device may be described as being used, but one of ordinary skill in the art will appreciate that the processing device includes a plurality of processing elements and / or a plurality of types of processing elements. It can be seen that it may include. For example, the processing device may include a plurality of processors or one processor and one controller. In addition, other processing configurations are possible, such as parallel processors.

The software may include a computer program, code, instructions, or a combination of one or more of the above, and may configure the processing device to operate as desired, or process independently or collectively. You can command the device. Software and / or data may be any type of machine, component, physical device, virtual equipment, computer storage medium or device in order to be interpreted by or to provide instructions or data to the processing device. It can be embodied in. The software may be distributed over networked computer systems so that they may be stored or executed in a distributed manner. The software and data may be stored on one or more computer readable recording media.

The method according to the embodiment may be embodied in the form of program instructions that can be executed by various computer means and recorded in a computer readable medium. The computer readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the media may be those specially designed and constructed for the purposes of the embodiments, or they may be of the kind well-known and available to those having skill in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical media such as CD-ROMs, DVDs, and magnetic disks, such as floppy disks. Magneto-optical media, and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like.

Although the embodiments have been described by the limited embodiments and the drawings as described above, various modifications and variations are possible to those skilled in the art from the above description. For example, the described techniques may be performed in a different order than the described method, and / or components of the described systems, structures, devices, circuits, etc. may be combined or combined in a different form than the described method, or other components. Or even if replaced or replaced by equivalents, an appropriate result can be achieved.

Therefore, other implementations, other embodiments, and equivalents to the claims are within the scope of the claims that follow.

Claims (7)

In the usage notification method performed by the usage notification system,
Analyzing and predicting a power consumption pattern through machine learning about power consumption; And
Providing estimated power rate information within a billing period associated with a progressive rate based on the analyzed and predicted power consumption pattern,
Providing the estimated power rate information within the billing period associated with the progressive rate based on the analyzed and predicted power consumption pattern,
And calculating a final billing history within a billing period in advance by adjusting the consumption of each device for the analysis and the predicted power consumption pattern through a simulation based on the following equation.
[Equation]

Where s is the electricity rate step, S is the electricity rate step corresponding to the total amount of power used,

Is the amount of power used within the s phase

),

Is the interval-based rate for step s of the electricity rate table (

),

Is the simulation result,

Is the maximum power limit for the interval in phase s of the electricity rate table

), Where Rs is the fare in the s phase,

Is the base rate, and R is the final charge amount.
delete The method of claim 1,
Providing the estimated power rate information within the billing period associated with the progressive rate based on the analyzed and predicted power consumption pattern,
Proposing a power consumption pattern for the remaining charging period of the charging period based on the predicted power consumption pattern
Usage notification method comprising a. The method of claim 1,
Analyzing and predicting a power consumption pattern through the machine learning about the power consumption,
Classifying power consumption information of each electronic device through the machine learning from total power consumption of electronic devices used in a specific environment;
Usage notification method comprising a. The method of claim 4, wherein
Analyzing and predicting a power consumption pattern through the machine learning about the power consumption,
Acquiring one or more of time-consuming consumption patterns from power consumption information for each electronic device or power consumption patterns from power consumption information for each electronic device;
Usage notification method comprising a. The method of claim 3,
Providing estimated power rate information within the billing period associated with the progressive rate based on the analyzed and predicted power consumption pattern,
The proposed power consumption pattern, the amount of power accumulated up to now from each electronic device, the total accumulated power amount expected within the charging period, the amount of power accumulated up to now from the electronic devices existing in a specific environment, or the estimated power rate information. Notifying the notification information
Usage notification method comprising a. In the usage notification system,
A prediction and analysis unit for analyzing and predicting a power consumption pattern through machine learning about power consumption; And
A provision unit for providing estimated power rate information within a charging period linked to a progressive rate based on the analyzed and predicted power consumption pattern,
The providing unit, the usage notification system, characterized in that to calculate in advance the final billing history within the billing period by adjusting the consumption per device for the analysis and predicted power consumption pattern based on the following equation.
[Equation]

Where s is the electricity rate step, S is the electricity rate step corresponding to the total amount of power used,

Is the amount of power used within the s phase

),

Is the interval-based rate for step s of the electricity rate table (

),

Is the simulation result,

Is the maximum power limit for the interval in phase s of the electricity rate table (

), Where Rs is the fare in the s phase,

Is the base rate, and R is the final charge amount.

Images