KR102472573B1 - Method of clustering and analyzing data of customer using electricity - Google Patents

Abstract

It relates to a cluster analysis method for electricity user data. The cluster analysis method for electricity user data includes the steps of extracting continuous data by extracting power consumption characteristics of electricity customers through time-of-use (TOU) indicators, and clustering the continuous data of the customers. , Clustering the categorical data of the customer, matching the clustering result of the continuous data and the clustering result of the categorical data to determine the final clustering, and determining the characteristics of the customer through the final clustering result. It includes an analysis step.

Description

Method for cluster analysis of electricity user data {METHOD OF CLUSTERING AND ANALYZING DATA OF CUSTOMER USING ELECTRICITY}

The present invention relates to a cluster analysis method for electricity user data, and more particularly, to a method for clustering and analyzing electricity user data including continuous data and categorical data.

An advanced metering infrastructure (AMI) is an interactive remote metering infrastructure that provides information to customers by measuring, collecting, and analyzing data on power consumption and quality in real time through smart meters. As various data are collected through AMI, various energy management services are being operated using it. In order to save energy and improve efficiency, energy management services must be operated according to the various characteristics of electricity users. Therefore, in order to provide an effective energy management service, a technique for rationally grouping customers according to characteristics using various data possessed by customers is required.

Domestic Registered Patent No. 10-1070368 (published date: 2011.03.07)

The problem to be solved by the present invention is to provide a method for analyzing electricity user data clusters that reasonably groups and analyzes customers according to characteristics using various data possessed by customers.

In order to achieve the problem to be solved, the method for analyzing the electricity user data cluster according to the embodiments of the present invention extracts the power consumption characteristics of the electricity user through a time-of-use (TOU) indicator and uses a continuous type Extracting data, clustering the continuous data of the customer, clustering the categorical data of the customer, matching the clustering result of the continuous data and the clustering result of the categorical data, and final clustering determining, and analyzing characteristics of the customer through the final clustering result.

According to embodiments of the present invention, it is possible to intuitively analyze a customer's power consumption characteristics using the TOU index. And since the number of clusters can be reasonably calculated through MCA, it is possible to effectively cluster and analyze mixed-type data of customers using electricity. In addition, by analyzing the data according to the cluster results, it is possible to propose an efficient energy management service operation plan according to the characteristics of electricity users in each cluster.

A detailed description of each drawing is provided in order to more fully understand the drawings cited in the detailed description of the present invention.
1 is a flowchart illustrating a method for analyzing a cluster of electricity user data according to an embodiment of the present invention.
2 is a block diagram illustrating a method for analyzing a cluster of electricity user data according to an embodiment of the present invention.
3 is a table showing TOU index and social factor data for each cluster according to a cluster result in the method for cluster analysis of electricity customer data according to an embodiment.
4 is a graph showing a three-dimensional scatter diagram through a TOU index as a result of clustering in a method for cluster analysis of electricity customer data according to an embodiment.

Specific structural or functional descriptions of the embodiments according to the concept of the present invention disclosed in this specification are only illustrated for the purpose of explaining the embodiments according to the concept of the present invention, and the embodiments according to the concept of the present invention may be embodied in many forms and are not limited to the embodiments described herein.

Embodiments according to the concept of the present invention can apply various changes and can have various forms, so the embodiments are illustrated in the drawings and described in detail in this specification. However, this is not intended to limit the embodiments according to the concept of the present invention to specific disclosed forms, and includes all modifications, equivalents, or substitutes included in the spirit and scope of the present invention.

Terms such as first or second may be used to describe various components, but the components should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another, e.g. without departing from the scope of rights according to the concept of the present invention, a first component may be termed a second component and similarly a second component may be termed a second component. A component may also be referred to as a first component.

It is understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, but other elements may exist in the middle. It should be. On the other hand, when a component is referred to as “directly connected” or “directly connected” to another component, it should be understood that no other component exists in the middle. Other expressions describing the relationship between components, such as "between" and "directly between" or "adjacent to" and "directly adjacent to", etc., should be interpreted similarly.

Terms used in this specification are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, terms such as "comprise" or "having" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in this specification, but one or more other features It should be understood that it does not preclude the possibility of the presence or addition of numbers, steps, operations, components, parts, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and unless explicitly defined in this specification, it should not be interpreted in an ideal or excessively formal meaning. don't

Hereinafter, a method for analyzing a cluster of electricity user data according to an embodiment of the present invention will be described.

1 is a flowchart illustrating a method for analyzing a cluster of electricity user data according to an embodiment of the present invention. The cluster analysis method for electricity user data may refer to a method performed by a computing device including at least a processor and/or a memory.

Referring to FIG. 1 , power consumption characteristics of customers using electricity may be extracted through TOU (Time-of-Use) indicators (S110). In this case, the data of customers using electricity may refer to data received through wired/wireless communication or received through a predetermined input device and then stored in the computing device.

The electricity customer data may include at least one of continuous data, categorical data, and mixed data. An example of continuous data is electric power data, and categorical data is customer's social factor data, such as current state of device ownership, number of household members, and region of residence. Mixed data is data that includes both continuous data and categorical data.

On the other hand, in the case of power data, since it is time-by-time (TOU) data measured in predetermined time units (eg, 15-minute units), the amount of data is large. Due to the characteristics of such power data, intuitive analysis is difficult during data analysis, and low clustering performance is shown even during cluster analysis. Therefore, there is a need to extract power consumption characteristics from power data for effective cluster analysis.

Electricity users need to extract the power consumption characteristics for each time zone because the power consumption varies by time zone according to their life pattern. Since power consumption is closely related to electricity rates, the characteristics can be intuitively displayed when extracting the customer's power consumption characteristics using the electricity rate system. According to an embodiment, among the electricity rates, a time-of-use (TOU) rate system that differentially charges electricity rates according to seasons and time zones exists. The TOU rate system divides 24 hours into three types: light load, medium load, and maximum load, and since each time zone is different depending on the season, the TOU indicator, which is an indicator using the TOU rate plan structure, determines the customer's electricity consumption according to the season and time zone. characteristics can be extracted.

The TOU index is an index representing the customer's power consumption characteristics according to seasons and time zones using the TOU rate structure, and is expressed according to Equation 1 below.

[Equation 1]

는 고객 'i'의 시간대별 전력 소비 특성(period = 경부하, 중간부하, 및 최대부하)이고,

는 시간대별 시간 가중치로 24시간 중 해당 시간대가 차지하는 비율이며,

는 고객 'i'의 해당 시간대 평균 전력이다.From here,

Is the power consumption characteristic of customer 'i' by time period (period = light load, medium load, and maximum load),

is the time weight of each time zone, which is the percentage occupied by that time zone in 24 hours,

is the average power of customer 'i' during that time period.

When clustering according to customer characteristics using electricity customer data, similarity between data can be calculated (S120). The similarity of data is analyzed by measuring the distance between data and can be calculated differently depending on the type of data. According to an embodiment, the similarity can be calculated with Euclidean distance for continuous data, the similarity can be calculated with Jaccard distance for categorical data, and the similarity can be calculated with Gower distance for mixed data. have.

Since the TOU index is continuous data, similarity can be calculated using the Euclidean distance. The Euclidean distance is a straight line distance between two points, and the Euclidean distance of the TOU index can be calculated according to Equation 2 below.

[Equation 2]

For customers using electricity, there is categorical data that shows social factors such as whether or not they have home appliances such as air conditioners and electric heaters, the total number of household members, and the region of residence. The similarity of categorical data is calculated by the Jacquard distance. The Jacquard distance is calculated using the union and intersection between the two sets S _i and S _j , as shown in Equation 3.

[Equation 3]

The similarity of mixed data can be calculated from customer to customer using distance. The distance between customers can be calculated as in Equation 4 below.

[Equation 4]

는 특성 'k'데이터 개수의 역수,

는 특성 'k'에서의 유사도로 데이터의 형태에 따라 다르게 계산될 수 있다. 두 고객 사이의 가워 거리가 가깝다는 것은 두 고객의 전체적인 특성이 유사하다는 것을 의미한다.From here,

is the reciprocal of the number of features 'k' data,

may be calculated differently according to the type of data with similarity in feature 'k'. A close distance between two customers means that the overall characteristics of the two customers are similar.

Since electricity customers need mixed data with power consumption characteristics and social factor data, the optimal number of clusters can be calculated by considering both types of data. The Matched Clustering Algorithm (MCA) is an algorithm for calculating the optimal number of clusters considering the shape of power consumption characteristics and social factor data.

Here, if the number of clusters is set too high in the data cluster analysis algorithm, there is an overfitting problem, so it is important to properly calculate the number of clusters. In general, the optimal number of clusters is calculated by graphing the Within Cluster Sum of Squares (WCSS), which is the sum of squares of distances between individuals belonging to each cluster, while increasing the number of clusters. And the point where the slope of the WCSS graph becomes gentle is calculated as the optimal number of clusters, and this method is called the elbow method. However, the elbow technique may appear differently depending on the type of data or the clustering algorithm, and since there is no clear criterion for smoothing the slope, it appears differently depending on the user. Therefore, according to an embodiment of the present invention, a matching clustering algorithm (MCA) is used to calculate a reasonable number of clusters by simultaneously reflecting various data types and clustering algorithms.

2 is a block diagram illustrating a method for analyzing a cluster of electricity user data according to an embodiment of the present invention.

Referring to FIGS. 1 and 2 , MCA may cluster power consumption characteristics and social factor data through K-means and Hierarchical Agglomerative Clustering (HAC), respectively (S130). At this time, the number of clusters is varied from a (where a is a natural number greater than 2 and an exemplary value may be 2) to b (where b is a natural number greater than a and an exemplary value may be 10). In addition, the number of clusters with the highest matching rate can be calculated as the optimal number of clusters by matching the K-means and HAC cluster results according to the number of clusters (S140).

The matching rate may be calculated based on the matching rate of the results of the two clustering algorithms. For example, the matching ratio between the clustering result according to the K-means algorithm (ie, the cluster number to which each customer belongs) and the clustering result according to the HAC algorithm (ie, the cluster number to which each customer belongs) can be calculated as the matching rate. . In FIG. 2, when the number of clusters is 2, customer 1 has two algorithms because the clustering result according to the K-means algorithm is the second cluster (cluster 1) and the clustering result according to the HAC algorithm is the first cluster (cluster 0). The clustering results according to do not match. Conversely, for customer 2, the clustering result according to the K-means algorithm is the second cluster (cluster 1) and the clustering result according to the HAC algorithm is also the second cluster (cluster 1), so the clustering results according to the two algorithms match. In this way, the ratio of the number of matched customers to the total number of customers may be the matching rate.

Additionally, it is possible to analyze the characteristics of customers using electricity by clustering mixed-type data through K-medoid using the calculated number of clusters (S150). In this case, when calculating the optimal number of clusters, if there are a number of clusters with the same matching rate, the total number of clusters with a smaller total distance can be selected as the optimal number of clusters by calculating the total Gower distance inside the cluster.

According to embodiments of the present invention, since the TOU indicator divides and expresses power data according to seasons and time zones, customers' power consumption characteristics can be intuitively analyzed. And since the number of clusters can be reasonably calculated through MCA, it is possible to effectively cluster and analyze mixed-type data of customers using electricity. In addition, by analyzing the data according to the cluster results, it is possible to propose an efficient energy management service operation plan according to the characteristics of electricity users in each cluster.

3 is a table showing TOU index and social factor data for each cluster according to clustering results in a method for cluster analysis of electricity user data according to an embodiment, and FIG. 4 is a table showing a cluster analysis method for electricity user data according to an embodiment. , It is a graph showing a 3-dimensional scatter plot through the TOU index of clustering results.

In FIG. 3 , when social factor data is analyzed based on the clustering results, cluster 0 is a cluster that has the characteristics of having 1 to 2 household members, and clusters 1 and 2 show the characteristics of having 3 to 4 household members. In addition, cluster 2 may show characteristics with a larger house size and more air conditioners than cluster 1.

Referring to FIGS. 3 and 4 , cluster 0 has a smaller overall power consumption than other clusters, and cluster 2 shows a characteristic of having a higher power consumption during peak load hours than cluster 1.

As a result, since the social factor characteristics of each cluster and power consumption by time can be easily identified, the electricity demand management company can provide customers with a service to reduce electricity bills by devising ways to save electricity according to the characteristics of each cluster. can

The device described above may be implemented as a hardware component, a software component, and/or a set of hardware components and software components. For example, devices and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), It may be implemented using one or more general purpose or special purpose computers, such as a Programmable Logic Unit (PLU), 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. A processing device may also access, store, manipulate, process, and generate data in response to execution of software. For convenience of understanding, there are cases in which one processing device is used, but those skilled in the art will understand 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 can include. For example, a processing device may include a plurality of processors or a processor and a controller. Also, other processing configurations are possible, such as a parallel processor.

Software may include a computer program, code, instructions, or a combination of one or more of these, and may configure a processing device to operate as desired or process independently or collectively. The device can be commanded. Software and/or data may be any tangible machine, component, physical device, virtual equipment, computer storage medium or device, intended to be interpreted by or to provide instructions or data to a processing device. , or may be permanently or temporarily embodied in the transmitted signal wave. Software may be distributed on networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer readable media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. Program commands recorded on the medium may be specially designed and configured for the embodiment or may be known and usable to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. - Includes hardware devices specially configured to store and execute program instructions, such as magneto-optical media, ROM, RAM, flash memory, etc. Examples of program instructions include high-level language codes that can be executed by a computer using an interpreter, as well as machine language codes such as those produced by a compiler. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

Although the present invention has been described with reference to the embodiments shown in the drawings, this is only exemplary, and those skilled in the art will understand that various modifications and equivalent other embodiments are possible therefrom. For example, the described techniques may be performed in an order different from the method described, and/or components of the described system, structure, device, circuit, etc. may be combined or combined in a different form than the method described, or other components may be used. Or even if it is replaced or substituted by equivalents, appropriate results can be achieved. Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the attached claims.

Claims (7)

A method of cluster analysis of electricity usage customer data performed in a computing device including at least one processor,
extracting continuous data by extracting power consumption characteristics of electricity customers through time-of-use (TOU) indicators;
clustering the customer's continuous data;
clustering the categorical data of the customer; and
Determining final clustering by matching a clustering result of the continuous data and a clustering result of the categorical data;
The continuous data is clustered through a K-means algorithm by calculating similarity with Euclidean distance,
The categorical data is clustered through a Hierarchical Agglomerative Clustering (HAC) algorithm by calculating similarity with Jaccard distance,
The step of clustering the continuous data of the customer and the step of clustering the categorical data of the customer are clustered while changing the number of clusters,
In the step of determining the final clustering, the number of clusters with the highest matching rate is calculated as the optimal number of clusters by matching the clustering results of the K-means algorithm and the HAC algorithm for each number of clusters;
When calculating the optimal number of clusters, if there is a number of clusters with the same matching rate, the total Gower distance inside the cluster is calculated and the number of clusters with the smallest total distance is selected as the optimal number of clusters.
A method for cluster analysis of electricity customer data. According to claim 1,
The time-of-day indicator is an indicator of the customer's power consumption characteristics according to the season and time zone based on the time-of-day rate plan structure that differentially charges electricity rates according to the season and time zone, and the following equation Method for analyzing customer data clustering.

(From here,

Is the power consumption characteristic of customer 'i' by time period (period = light load, medium load, and maximum load),

is the time weight of each time zone, which is the percentage occupied by that time zone in 24 hours,

is the average power of customer 'i' during that time period.) delete delete delete A method for analyzing a cluster of electricity usage customer data performed in a computing device including at least one processor,
extracting continuous data by extracting power consumption characteristics of electricity customers through time-of-use (TOU) indicators;
clustering the customer's continuous data;
clustering the categorical data of the customer;
determining final clustering by matching a clustering result of the continuous data and a clustering result of the categorical data; and
Analyzing the characteristics of the customer through the final clustering result,
The mixed data including the continuous and categorical data is clustered through K-medoid by calculating the similarity with the Gower distance,
A method for cluster analysis of electricity customer data. delete

Images