See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/361841374

A STUDY OF THE MODELING ON THE SMART FACTORY PRODUCTION

OPTIMIZATION USING ENERGY CONSUMPTION PREDICTION

Article in International Journal of Advanced Science and Technology · January 2022

DOI: 10.33832/ijast.2022.150.04

CITATION READS

1 258

5 authors, including:

Rahman A.B.M. Salman Rahman Moonyong Lee

Sunchon National University Yeungnam University
16 PUBLICATIONS 60 CITATIONS 873 PUBLICATIONS 15,627 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Rahman A.B.M. Salman Rahman on 08 July 2022.

The user has requested enhancement of the downloaded file.

 International Journal of Advanced Science and Technology
Vol.150 (2022), pp.31-40
http://dx.doi.org/10.33832/ijast.2022.150.04

A STUDY OF THE MODELING ON THE SMART

FACTORY PRODUCTION OPTIMIZATION USING
ENERGY CONSUMPTION PREDICTION
^

A.B.M.Salman Rahman1, Myeongbae Lee2, Jonghyun Lim3, Yongyun Cho4 and

Changsun Shin4*
1,3
Student, Department of Information & Communication Engineering, Sunchon
National University, South Korea
2
Researcher, Department of Information & Communication Engineering, Sunchon
National University, South Korea
4
Professor, Department of Information & Communication Engineering, Sunchon
National University, South Korea
1
salman01717@gmail.com, 2lmb@scnu.ac.k, 3sshb56@s.scnu.ac.kr,
4
yycho@sunchon.ac.kr, 5csshin@sunchon.ac.kr

Abstract— Energy consumption is a global issue that significantly impacts the

nation's energy security. In the industrial sector, smart factory energy consumption
analysis and forecasts are crucial for improving energy usage rates and creating
profits. An industrial enterprise's energy analysis and forecasting are becoming
progressively essential. As a result, this is an excellent time to lend a helping hand
technologically to minimize energy consumption and deliver profitable outcomes for
smart factories. It is really challenging to evaluate energy usage and make reliable
estimations of industrial energy consumption. In a nutshell, this paper examines
monthly electricity consumption to identify the discrepancy between energy
utilization and energy demands. It represents the connection between energy
consumption, energy demand, and various manufacturing goods on a graph. This
study compares real data with forecasted data curves using the Autoregressive
integrated moving average (ARIMA), and Seasonal Autoregressive integrated
moving average(SARIMA) to improve energy usage. ARIMA and SARIMA have the
Root Mean Square Error (RMSE) performance evaluations of 8.90 and 10.90,
respectively. Finally, determine the most effective product for the smart factory to
enhance the energy utilization rate and generate profit for the smart factory.

Keywords— Smart Factory, Energy Consumption, Correlation, Moving Average,

Data Forecasting, ARIMA, Seasonal ARIMA, Variable Importance

1. INTRODUCTION
In today's world, so much energy is consumed for economic and population expansion.
So, we must have to consider how we might minimize our energy use. Rapid economic
growth has increased global electric power and energy consumption. Meanwhile,
electricity is among the most significant drivers of economic development and is
indispensable in our everyday lives [1]. The industry uses electricity to process,
manufacture, or assemble products in a variety of manufacturing, mining, agriculture, and

^
Received: November 11, 2021
Reviewed: January 6, 2022
Accepted: January 17, 2022
*
Corresponding Author

ISSN: 2207-6360 IJAST

Copyright ⓒ 2022 NADIA
International Journal of Advanced Science and Technology
Vol.150 (2022)

construction to the advantage of customers and equipment. Overall, the industrial sector's
energy usees is more than a third of the nation's electricity. As a result, forecasting electrical
energy consumption has become essential for a country or area[2].
A Smart Factory is an Internet of Things concept that sees a manufacturing procedure
as a completely advanced integrated and smart network of technology that allows
infrastructure, types of equipment, and operational chains to run without human
intervention [3]. All of these things happen in a smart factory as a result of data interchange
between all elements in the production technology chain, not only between production tools
and equipment. Life has become more accessible thanks to the accomplishments of the
third scientific and technological revolution. It has pushed for new technology and
structural change in traditional industries. To further develop a modern processing plant's
energy usage, it is about to work with the leading administration group of mechanical
industrial facilities as a solid, innovative hand to develop their energy use [4] [5] further.
This study deals with analyzing usages of energy consumption for every month with
every hour data to find out the difference or relation between usages of energy consumption
and demands of energy consumptions. Make a relation curve among usages of energy,
demands of energy, and different types of materials weight for manufacturing product. To
find out the high significance relation between usages of energy and other materials weight
for manufacturing product using the correlation method. We forecasted energy
consumption using ARIMA and Seasonal ARIMA and compared the actual data curves
with predicted data curves to improve the energy utilization rate. Finally, determine the
most effective product for the smart factory to enhance energy utilization and obtain
lucrative items for the smart factory.

2. RELATED WORKS
Many scholars have studied the topic of estimating energy demand using data-mining
techniques for a long time. Several analytical and machine intelligence methods have been
developed to estimate energy demand patterns. Following the construction of a model, ML
algorithms are extraordinarily advantageous and comfortable for a typical operator to use;
they are becoming increasingly prevalent in a variety of applications [6]
Recently, scientists have concentrated on the utilization of Machine learning methods
for foreseeing energy utilization. Numerous factual and artificial reasoning procedures for
displaying energy utilization designs have been created. AI calculations are beneficial and
straightforward to a typical administrator after the model has been set up and predominant
in numerous applications [7]. Decreasing the utilization of power in the steel business is a
worldwide issue where moves are made energetically by the public authority. A smart
factory can enhance energy use control by demonstrating and determining that the request
is functional [8]. Energy preservation is necessary since the three significant financial
businesses, assembling, transport, and advancement, have weighty energy utilization. The
warming burden computation is the initial step of the development cycle for the HVAC
framework [9].
Electrical hardware and apparatus were the biggest supporters of the general
construction and strength impact and crude substance materials and compound parts sub-
areas, while the littlest commitments were from the gas and oil refining and coking
assembling and supply ventures[10]. Mechanical advancement has given tremendous
freedoms to specialists in assorted fields to utilize artificial brainpower. Different
endeavors have been made in the assembling and advancement fields to use AI strategies
[11].

32 Copyright ⓒ 2022 NADIA

 International Journal of Advanced Science and Technology
Vol.150 (2022)

3. METHODOLOGY
3.1. TIME SERIES FORECASTING
A time series is a collection of really well data points recorded at regular intervals over
a period of time. By fitting appropriate models to time series data points, time series
analysis helps us understand the underlying processes that contribute to a specific pattern
in time series data points and forecast and monitor the data points. Time series analysis is
particularly useful for analyzing monthly, seasonal, and yearly trends in sales. This study
uses two types of time series forecasting models: ARIMA and seasonal ARIMA
(SARIMA).

3.1.1 ARIMA FORECASTING: Attach ARIMA models are a well-known and

adaptable course of determining models that utilize chronicled data to create the forecast.
The auto-regressive integrated moving average (ARIMA) is defined by three order
parameters (p, d, q). The Box-Jenkins approach is a term used to describe the method of
fitting an ARIMA display. An auto-regressive component relates to past values in the
regression equation for the series Y. The auto-regressive parameter p specifies the number
of lags employed in the model.

3.1.2 SEASONAL ARIMA FORECASTING: Seasonal ARIMA is a seasonal

structure that may be used to specify an ARIMA model (SARIMA). This example describes
the model by two sets of order parameters (p, d, q) as order parameters and (P, D, Q) m as
model parameters. Parameters describe the seasonal component of m periods.

4. RECORDED DATA DESCRIPTION

In this study, data are collected from Daewoo Steel Factory, South Korea. Daewoo Steel
Factory is a well-known steel manufacturer company across the world. There are two types
of data sets available. One of these smart factory energy usages, while the other is for
manufacturing or production data. From energy data sets, we get twenty-four hours usages
of energy, demands of energy, lagging reactive power of the current, leading reactive power
of the current, percentage of leading power, and the percentage of lagging powers available.
From productions data, we get details about manufacturing like date of manufacturing
product, product name, materials length, materials weight, product item, and lots of
information. From manufacturing, there are three types of products manufactured in that
steel factory, and the products are "Sheet," "Skelp," and "Cyong." Table I presents different
types of load and their timings concerning each month. The table information is collected
from KEPCO.

Table I. Load Type and its Timings

Charged with Energy (won/kWh)
Charge on
Classification Demand Time Period Summer Spring/Fall Winter
(won/kW) (Jul.1~Aug31) (Mar.1~Jun.30/ (Nov.1~Feb28.)
Sep.1~Oct.31)
Off Peak load 45.80 45.80 49.30
Option
High 5710 Mid Load 87.30 59.30 85.90
I
voltage Peak Load 147.00 77.60 120.70
(A)
Off Peak load 41.60 41.60 45.20
Option Mid Load 83.20 55.20 81.70
6540
II
Peak Load
142.90 73.50 116.60

Copyright ⓒ 2022 NADIA 33

International Journal of Advanced Science and Technology
Vol.150 (2022)

5. RESULT
Based on two data sets, we analyze the total energy consumption of the year 2017, then
analyze per-month energy consumption for the year 2017 and make a comparison graph
between usages of energy and demands of energy in the year 2017. Analysis usages of
energy data with different material weight and make a comparison graph of total usages of
energy and total demands of energy with corresponding total materials weight. The
industry's energy(kWh) data for every 1-hour is the focus of this analysis. The 1-hour
reporting frequency was employed to capture the substantial fluctuations in energy use. All
the data analysis is carried out in R [12]. The timespan of data is 365 days (12 months). We
forecasted energy consumption using ARIMA and Seasonal ARIMA and compared the
actual data curves with predicted data curves to improve the energy utilization rate. Finally,
determine the most effective product for the smart factory to enhance energy utilization
and obtain lucrative items for the smart factory.

5.1. ANALYZING PER-HOUR ENERGY DATA

We examine total energy consumption per hour using 2017 energy consumption
statistics to estimate every hour's energy usages and establish the time periods with the
highest and lowest. Figure 1 depicts the energy use for working days in 2017 over a twenty-
four-hour period. The x-axis depicts the time in hours, while the y-axis depicts the total
energy use in kWh.

Fig. 1 The Diagram Shows the Per-Hour Usages of Energy for a Steel Factory in 2017
in the Box Plot

5.2 ANALYZING ENERGY DATA WITH MATERIALS WEIGHT OF

MANUFACTURING PRODUCTS:
As we told before, three types of products are manufactured in Daewoo Steel Factory.
Here we analyze different materials weight of manufacturing products with total usages of
energy in terms of the number of days on the job. Figure 2 represents different types of
material weight of manufactured products with per-day total usages of energy for the
number of working days. In 2017, the x-axis represented the number of working days, while
the y-axis represented the material weight in kilos. The blue line depicts daily energy use
for working days, while the red dots depict daily material weight for sheet productions,
green dots depict daily material weight for skelp products, and the purple line depicts daily
material weight for Cyeong productions.

34 Copyright ⓒ 2022 NADIA

 International Journal of Advanced Science and Technology
Vol.150 (2022)

Fig. 2 Different Type's Material Weight of Manufactured Products with Usages of

Energy for the year2017

Fig. 3 Materials Weight of Manufactured Products with Usages of Energy for the Year
2017

Figure 3 shows the everyday total material weight of manufactured products with total
usages and demands of energy consumptions for the number of working days. The x-axis
depicts working days, while the y-axis depicts the material weight in kilograms. The blue
line depicts energy usages for working days, the red line depicts a maximum demand for
energy, and the green line depicts the total weight of materials for the period of working
days, as shown in the figure.

5.3 ARIMA AND SARIMA FORECASTING

ARIMA and SARIMA are time series models that consist of a series of well-defined
data points taken at regular intervals over a length of time. ARIMA and SARIMA assist us
in understanding the underlying processes that contribute to a certain pattern in data points
and forecast and monitor those data points.

Copyright ⓒ 2022 NADIA 35

International Journal of Advanced Science and Technology
Vol.150 (2022)

Fig. 4 Forecasting of Energy Consumption of January 2018 based on 2017 Energy Data
by using ARIMA (7,1,4)order

Fig. 5 Forecasting of Energy Consumption of January 2018 based on 2017 Energy Data
by Using Seasonal ARIMA(7,1,4)(0,0,0)[30]order

Figure 4 shows the forecasted energy data based on the past 2017th year to predict the
January month of 2018 using the ARIMA Model in the order of (7,1,4). The X-axis has the
number of months in 2017, and the Y-axis has energy consumption kWh. The black color
line refers to actual energy consumption, the blue line refers to predicted energy
consumption, and the grey line refers to the maximum and minimum range of estimation
prediction. This plot is based on the past 2017 year data to predict the January month of
2018 energy consumption. Figure 5 shows forecasted energy data based on the past 2017th
year to predict the January month of 2018 by using the SARIMA Model in the order of
(7,1,4) (0,0,1) [30]. The X-axis has the number of months in 2017, and the Y-axis has
energy consumption kWh. The black color line refers to actual energy consumption, the
blue line refers to predicted energy consumption, and the grey line refers to the maximum
and minimum range of estimation prediction. This plot is based on the past 2017 year data
to predict the January month of 2018th energy consumption. The past and predicted data
are more connected because both have more fluctuation (variations).

36 Copyright ⓒ 2022 NADIA

 International Journal of Advanced Science and Technology
Vol.150 (2022)

Fig. 6 The Comparison between Actual Energy Data with ARIMA Forecasted Energy
Data Jan-2018

Figure 6 depicts a comparison between actual and anticipated energy usage statistics.
The X-axis represents the number of days between January 1 and January 31, 2018, while
the y-axis represents energy usage in kWh. The blue line represents ARIMA Model-
predicted data, and the pink line represents actual usages of energy. The green dotted line
represents the maximum and lower range of 80 percent confidence intervals. The gray
dotted line represents the top and lower range of 95 percent confidence intervals. So, based
on the predicted data, we suggest taking the upper range of 95 percent confident intervals
and the lower range of 80 percent confident intervals.

Fig. 7 The Comparison between Actual Energy Data with SARIMA Forecasted Energy
Data Jan-2018

Figure 7 shows the comparison of original energy consumption data and forecasted
energy consumption data. The X-axis has the number of days from 1 to January 31, 2018,
and the Y-axis has energy consumption kWh. The blue line refers to predicted data using
SARIMA Model, and the pink line refers to actual energy consumption. The green dotted
line refers to the upper & lower range of 80% confident intervals. The gray dotted line
refers to the upper & lower range of 95% confident intervals. So, based on the predicted
data, we suggest taking the upper range of 95% confident intervals and the lower range of
80% confident intervals.

Copyright ⓒ 2022 NADIA 37

International Journal of Advanced Science and Technology
Vol.150 (2022)

Table II. Forecasting Models Performance for Daewoo Steel Factory Data Set for
ARIMA & SARIMA
Model ME RMSE MAE
ARIMA .59 8.70 6.72
SARIMA .24 10.90 8.79

The Daewoo steelfactory manufactures a variety of energy-related goods, and

manufacturing products are in the name of Skelp, Sheet, Cyong, and packaging. Table III
shows the using percentage of energy consumption for each manufacturing product. From
the table, we find out that Skelp uses 49.82 percent of energy from total uses of energy.

Table III. Daewoo Steel Factory Energy Consumption Table for the Year 2017
Product Usages of energy (kWh) Using Percentage
Name
Sheet 322522.80 39.70%
Skelp 404568.10 49.82%
Cyong 45508.81 5.62%
Packaging 34836.55 4.32%
Others 807436.97 99.46%

6. DISCUSSION
All of the findings and analysis revealed a clear link between energy use and the varying
material weights of manufactured goods. We can see from figure 1 that industrial energy
consumption is high from 8 a.m. to 10 p.m. every working day, and the energy consumption
is low from 11 p.m. to 8 a.m. From Figure 2, we can say that energy consumption also
increases when the density of materials increases. It means when productions are high,
usages of energy consumption also high. Figure 4 & 5 shows the forecasted energy data
based on the past 2017th year to predict January 2018 using the ARIMA and SARIMA
Model. Figure 6 & 7 compares original energy consumption data and forecasted energy
consumption data by using ARIMA and SARIMA model. From Table II, we can easily
find the performance of ARIMA and SARIMA forecasting models. Table II shows the
forecasting model performance for the Daewoo Steel factory for ARIMA and SARIMA.
From the table, we can find the performance evaluation result for ARIMA by M.E., RMSE
& MAE are .59, 8.90, and 6.72. Performance evaluation results for SARIMA by M.E.,
RMSE, & MAE are .24, 10.90, 8.79, and Table III shows that Skelp uses 49.82 percent of
energy from total uses of energy. From Table III, we can easily say Skelp is the most
effective product for the Daewoo Steel Factory.

7. CONCLUSIONS
This study focuses on industrial energy usage analysis and forecasting to enhance energy
utilization rates and make steel plants more profitable. This study shows the one-month
forecasting curve for steel factories using the ARIMA and SARIMA time series models
and finds the comparison curve between the actual and forecasting values. Table III shows
that most energy is used for Skelp production at percent 49.82, Sheet production at 39.70
percent, and Cyong production at 5.62 percent. In 2017, this smart factory produced the
most Skelp with the maximum amount of electricity. So, to make energy a lucrative item
for the business, we propose to operate two shifts in the factory, one during the day and
one during the night, such as during peak hours and off-peak hours, so that the firm's energy

38 Copyright ⓒ 2022 NADIA

 International Journal of Advanced Science and Technology
Vol.150 (2022)

costs will reduce. As we have discovered, skelp production is closely connected with
energy use, so allow authorities to manufacture Skelp during the night shift.

ACKNOWLEDGMENTS
The Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant
funded by the Korean government (MOTIE) (20202020900060) have been sponsored this
work. This research has been funded by the MSIT (Ministry of Science and ICT), Korea,
under the Grand Information Technology Research Center support program (IITP 2021-
2020-0-01489) supervised by the Institute for Information & communications Technology
Planning & Evaluation (IITP). This research was supported by the Interdisciplinary
Program in IT-Bio Convergence System, Sunchon National University, 255, Jungang-ro,
Suncheon-si, Jeollanam-do 57922, Republic of Korea.

REFERENCES
[1] Xiao, L., Shao, W., Liang, T. and Wang, C., “A combined model based on multiple seasonal patterns and
modified firefly algorithm for electrical load forecasting”, Applied Energy, vol. 167, no. C, (2016), pp.
135-153.
[2] Wang, Z.-X., Li, Q. and Pei, L.-L., “A seasonal G.M. (1, 1) model for forecasting the electricity
consumption of the primary economic sectors”, Energy, vol. 154, (2018), pp. 522-534.
[3] Salman Rahman, A. B. M., Lee, M., Lim, J., Cho, Y. and Shin, C., “A Comparative Study of Energy Big
Data Analysis for Product Management in a Smart Factory”, Journal of Organizational and End User
Computing, (2022) DOI: 10.4018/JOEUC.291559.
[4] Salman Rahman, A. B. M., Lee, M., Lim, J., Cho, Y. and Shin, C., “A prediction model for steel factory
manufacturing product based on energy consumption using data mining technique”, Journal of Science
and Engineering Management, vol. 1, no. 2, (2020), pp. 9-16.
[5] Salman Rahman, A. B. M., Lee, M., Ragu, V., Lim, J., Cho, Y., Park, J. and Shin, C., “Identification of
High Significance Product Items Through the Analysis of Energy Consumption in Steel Factory”, Journal
of Knowledge Information Technology and Systems(JKITS), vol. 14, no. 3, (2019), pp. 275-289.
[6] Sathishkumar, V. E., Shin, C. and Cho, Y., “Efficient energy consumption prediction model for a data
analytics-enabled industry building in a smart city”, Building Research & Information, (2020) DOI:
10.1080/09613218.2020.1809983.
[7] Simeone, O., “A very brief introduction to machine learning with applications to communication systems”,
IEEE Transactions on Cognitive Communications and Networking, vol. 4, no. 4, (2018), pp. 648-664.
[8] Chen, C., Liu, Y., Kumar, M., Qin, J. and Ren, Y., “Energy consumption modeling using deep learning
embedded semi-supervised learning”, Computers & Industrial Engineering, vol. 135, (2019), pp. 757-765.
[9] Chou, J. S. and Bui, D. K., “Modeling heating and cooling loads by artificial intelligence for energy-
efficient building design”, Energy and Buildings, vol. 82, (2014), pp. 437-446.
[10] Zha, D., Zhou, D. and Ding, N., “The contribution degree of sub-sectors to structure effect and intensity
effects on industry energy intensity in China from 1993 to 2003”, Renew. Sustain. Energy Rev., vol. 13,
(2009), pp. 895-902.
[11] Paturi, U. M. R. and Cheruku, S., “Application and performance of machine learning techniques in the
manufacturing sector from the past two decades: A review”, Materials Today: Proceedings, vol. 38,
(2020), pp. 2392-2401.
[12] Team, R. C., “R-A language and environment for statistical computing”, Vienna: R Foundation for
Statistical Computing; 2014 [Zugriff/access Jun 2018]. Verfügbar unter.

Copyright ⓒ 2022 NADIA 39

 International Journal of Advanced Science and Technology
Vol.150 (2022)

40 Copyright ⓒ 2022 NADIA

View publication stats