Scribd
Upload
10 views

Final Report

The paper presents a k-means clustering method to analyze residential electricity customers' load profiles using extreme points and demographic characteristics. It identifies six distinct energy consumption patterns among 2041 households, which can aid in developing personalized electricity market strategies. The authors suggest that real-time data integration and advanced forecasting methods could enhance the clustering process for better market efficiency.

Uploaded by

Nadya Noorfatima
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
10 views

Final Report

The paper presents a k-means clustering method to analyze residential electricity customers' load profiles using extreme points and demographic characteristics. It identifies six distinct energy consumption patterns among 2041 households, which can aid in developing personalized electricity market strategies. The authors suggest that real-time data integration and advanced forecasting methods could enhance the clustering process for better market efficiency.

Uploaded by

Nadya Noorfatima
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 4

21/12/202

Research Paper Reading Sheet Date


1

Clustering of Load Profiles of Residential Customers Using Extreme


Title
Points and Demographic Characteristics

Author Hyun Cheol Jeong, Minseok Jang, Taegon Kim, Sung-Kwan Joo

Published in
(Conference/journal, Year)
Electronics, MDPI, 26/01/2021

Keyword

Extreme points, demographic characteristics, customer load grouping

Contents (Name of the sections in the paper)

I. Introduction
The level of electrification has been increased significantly in the past few years. Furthermore,
the types of electricity customers also have been varied significantly differentiated by their
demographic profile. In this case, the analysis of the customer load profile can be related to
marketing and design of the load management program. This paper proposes k-means clustering
method for electricity consumers by implementing extreme points and feature selection as the
clustering parameters. The clustering analysis of energy consumption customers is carried out
using load profiles from 2041 households with various demographic characteristics is presented.

II. Selection of Extreme Points of The Energy Consumption Pattern with Demographic
Characteristics
For the selection of extreme points, initially all the data is evaluated by using z-score to detect
outliers. By doing this, the clustering result accuracy will be increased. After deleting outliers,
the load profile is normalized to determine the extreme points: morning local minimum point,
morning local maximum point, evening local minimum point, and evening local maximum point.
Furthermore, the last step of this section is feature selection. Feature selection is undertaken by
analyzing the correlation between monthly total energy consumption and customer demographic
data: area, number of household members, salary level, age, and education level.

III. Clustering Method for Residential Customer Energy Profile Using Extreme Points and
Demographic Characteristics
In this section, the k-means clustering method is performed based on the selected features and
extreme points. For the load profiles of residential customers, the difference in energy
consumption between customers is small and the range of energy consumption is wide. In this
section, clustering of load profiles of residential customers is carried out by computing the
distance from the cluster center point in the residential customer energy consumption data. The
load profiles of residential customers are clustered by minimizing the sum of the Euclidean
distances of the group set. The elbow method is adopted to compute the distance from the cluster
center point to residential customer load profile in order to determine the optimal number of
clusters.

IV. Results
According to the k-means clustering with determined extreme points and selected features
results, there are six categories of patterns of residential customers. The patterns of residential
customers are clustered by utilizing K-means into energy consuming, energy saving, M-pattern,
morning concentrated, evening concentrated, and owl. Energy consuming is a typical energy
consumption pattern for residential customers. The energy saving cluster is a type of consumer
that saves energy in the morning. The M-pattern cluster with similar energy consumption in the
morning and the evening, such as office workers which do not stay at home during the day.
Morning concentrated cluster is a type of customer with energy consumption concentrated in the
morning. The evening concentrated is type of energy consumers with energy consumption
concentrated in the evening. Owl cluster has relatively large energy consumption at daybreak
compared to the other clusters with very low energy consumption during the daytime.

Summary of Paper (in your own words)-Do not copy from the abstract

Electricity customer has a very wide variations, that brings difficulty for the utility company to provide
optimal services through electricity market. To help analyze residential customers for constructing the
electricity market strategies, the clustering of load profiles by using extreme points and demographic
characteristics. The results show that the residential customers can be categorized into six types of energy
consumption patterns, and the total monthly energy consumption, area, and number of household members
are analyzed in each cluster. This research then can be a promising solution to be directed toward the
development of personalized residential electricity rates for different demographic characteristics of
customers.

Limitation/Extension (Weakness of the proposed idea, can you think of any improvement?)

The limitation of this work: The load profile analyzed for this work is based on historical data which the
clustering is done on one time only. Meanwhile, for the electricity market, the data should be real
time in order to maintain the proper strategy. Such as, the load balancing process in power system
operation will be determined based on the electricity market operation. In addition, the clusters
determination has significant effect towards electricity market strategy. Thus, if the k-means
clustering is used as the proposed method, for a real-time result, machine learning can be used to do
forecasting such as short term load forecasting by using Long Short-Term Memory (LSTM) and
convolutional neural network (CNN). Then, by implementing short term load forecasting, the
clustering can be performed in real time. In this case, the clustering of load profiles should be
improved by considering real time aspect such as dynamic electricity tariff, real time load capacity,
weather condition, etc. To give another perspective, the figure below presents the flowchart of
paper’s improvement.
Your Opinion

Explain why you think this is a good or bad idea.

The idea to cluster residential customers is definitely a


breakthrough that can only happen due to the development of
Quality of the idea (Do not just unsupervised learning. Due to the implementation of Advanced
write ‘good’ or ‘bad’. Explain) Metering Infrastructure (AMI) as well, the customer’s data
collection is possible to performed in real time. Due to the large
amount of data is available, the clustering method helps to
analyze the measured data, including load profile in order to be
used on the future electricity market development.

Potential impact (How is the


proposed idea going to make a
difference in research or in Explain why this idea can have some impact and how?
practice?)
Clustering of electricity customer gives new perspective in the
development of electricity market. The major challenge of
development of electricity market is to provide proper market
strategy to maintain fair profit allocation among participants. Fair
profit allocation is difficult to obtain since the wide variations of
electricity customers even in residential customers. Therefore,
clustering of electricity customer is going to be impactful to
provide personalized electricity market which is constructed
based on the selected features. For the features itself, it can be
derived into a wider topic to develop in order to improve the
clustering results.

In practical, electricity market based on clustering process can be


performed by using discriminatory auction mechanism or widely
known as pay-as-bid electricity market. Pay-as-bid electricity
market allows market participants with various electricity tariff
and consumption capacity. However, this type of market
mechanism can’t guarantee efficient market results, since the
resulted electricity tariff also has wide range of market prices. In
this case, by using clustering method, the participants are grouped
and then pay as bid electricity market can be performed in more
efficient manner.

You might also like