Machine Learning Decision Support Model for Greenhouse Gas

Reduction Technology Application

I. M. Aliev

Chechen State University, Grozny, Russia

Keywords: Reduction technology, machine learning, GBRT, SVM, DNN, K-ETS.

Abstract: Many countries have implemented policies to reduce greenhouse gas (GHG) emissions since the 21st

Conference of the Parties (COP 21) of the United Nations Framework Convention on Climate Change

(UNFCCC) in 2015. The parties to this convention have voluntarily agreed to a new climate regime that aims

to reduce greenhouse gas emissions. Subsequently, reducing greenhouse gas emissions through specific

reduction technologies (renewable energy) to reduce energy consumption has become a necessity rather than

an option. With the launch of the Korea Emissions Trading Scheme (K-ETS) in 2015, Korea has certified and

funded projects to reduce greenhouse gas emissions. To help the user make informed decisions about the

economic and environmental benefits of using renewable energy, an evaluation model has been developed.

This study establishes a simple assessment method (SAM), an assessment database (DB) of 1199 greenhouse

gas reduction technologies implemented in Korea, and a machine learning-based greenhouse gas reduction

technology assessment model (GRTM). In addition, proposals are made to assess the economic benefits that

can be obtained in combination with the environmental benefits of technology to reduce greenhouse gas

emissions.

1 INTRODUCTION

After the lead version of the Kyoto Protocol of the

United Nations Framework Convention on Climate

Change. In 1997, many countries around the world,

including Korea, made significant efforts to reduce

greenhouse gas (GHG) emissions. In particular,

building energy efficiency has been highlighted both

in Europe and abroad, and many countries are

strengthening building energy efficiency policies by

regulating building design standards, promoting

environmentally friendly benefits, and introducing

zero-energy buildings (Korea Energy Agency, 2018).

In order to successfully reduce spending across the

country, led by the public sector, Korea included an

incentive system for a new climate regime under 100

policy objectives. This required the republican

institution to maintain membership fees (Ministry of

Environment, 2019). As a result, starting in 2020, any

newly constructed building owned by public

institutions is required to introduce energy efficiency

and achieve a 30% reduction in baseline emissions (in

accordance with the agreements of 2007, 2008 and

2009) by 2030 (Ministry of Environment, 2017).

However, lack of experience and understanding

presents a major challenge to the implementation of

this system and is the main reason why the policy is

not effective. Several studies have been carried out to

develop policies to reduce energy consumption or to

analyze the implications of reducing greenhouse gas

emissions from renewable energy. In addition, using

the Emissions Trading Scheme (ETS), a system

designed to support emission reductions, government

agencies voluntarily undertake external emission

reduction projects, certify emission reductions, and

receive economic benefits from emissions trading

(Ministry of Environment, 2018). In previous studies

on the efficiency and control of greenhouse gas

emissions, a greenhouse gas emission prediction

system was developed, which determines the factors

that contribute to emissions fluctuations, with an

emission control system that compares different

buildings based on the developed prediction model. It

is noteworthy that Che (2017) used data analysis

technology to increase the renewable energy market

and compare indicators to select the optimal building

for simulation (Chan, 2017). Chan (2018) predicted

energy production from real-time energy generation

database based on machine learning (Jang, 2018).

Aliev, I.

Machine Learning Decision Support Model for Greenhouse Gas Reduction Technology Application.

DOI: 10.5220/0011556700003524

In Proceedings of the 1st International Conference on Methods, Models, Technologies for Sustainable Development (MMTGE 2022) - Agroclimatic Projects and Carbon Neutrality, pages

165-168

ISBN: 978-989-758-608-8

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

165

2 MATERIALS AND METHODS

A study by nakawa (2018) on reducing energy

consumption and reducing greenhouse gas emissions

analyzed the impact of replacing conventional lamps

with LEDs. Singh (2015) proposed GHG emission

reduction guidelines for GHG emissions forecast up

to 2030 under the ezp program for Israel (Nakano,

2018; Singh, 2015). Further analysis of the efficiency

of renewable energy sources is needed. In addition,

national policy, along with economic and

environmental decision-making regarding renewable

energy sources, should be changed based on the

results of a comprehensive assessment of the energy

performance of buildings. Since the World Economic

Forum in Davos, Switzerland in January 2016, the

idea of a fourth industrial revolution has become a

controversial issue around the world. The fourth

industrial revolution refers to a consumer-oriented

economy based on technological automation and

represents an era when the world is understood

through artificial intelligence (AI) (Bottou). AI refers

to technology that analyzes data through

visualization, machine learning, and data mining to

find better solutions. In AI, machine learning is a

technique that reads huge amounts of data, finds an

algorithm, and predicts changes. In the field of

environmental architecture, there is a growing

number of cases where machine learning is applied to

a system for predicting economic and environmental

benefits. When comparing the predictive ability

between models using traditional statistical technique

and models using machine learning technique,

various studies have shown that the model using

machine learning has a higher level of predictive

ability (Jordan). Therefore, this study proposes an

optimal machine learning model for GHG projects

with a high level of predictive power that uses real

renewable energy sources and examples of external

GHG projects. We believe that this model is superior

to costly and inefficient methods such as energy

modeling. Through this work, we aim to help select

projects with the most significant economic and

environmental benefits compared to the actual energy

consumption of an existing building.

This can be achieved by looking at energy

production in greenhouse gas emission reduction

projects that cover renewable energy, installation

cost, efficiency and energy consumption by use and

source, all of which contribute to emission reductions.

A database of certified GHG emission reductions

from renewable energy sources was created and used

as the basis for building a machine learning model to

support decision-making on GHG emission reduction

projects. To analyze the predictive power of the fitted

model, we used principal component analysis (PCA),

selected gradient boosted regression tree (GBRT),

support vector machine (SVM), and deep neural

network (DNN) machine learning algorithms

optimized for the original data set. in this study. We

also tested the predictive power of machine learning

methods by comparing their predictive power with a

multiple regression model. Unlike previous studies

that only compared the predictive power of different

models, this study examines the applicability of

machine learning to greenhouse gas reduction

technologies by analyzing calculated numerical

values. The Korean government subsidizes a certain

portion of the installation costs to promote the spread

of renewable energy. In response, businesses are

adopting government-subsidized renewable energy

sources such as solar power, solar thermal power, and

geothermal power, along with efficient technologies

including high-efficiency lighting, LED lighting, and

green vehicles (Korea Energy Agency, 2018).

To the extent that reduced energy consumption in

green vehicles can be converted into economic

benefits, this study includes green vehicles in

greenhouse gas emission reduction technologies. We

have defined the introduction of renewable energy

and high-efficiency equipment as "technologies to

reduce greenhouse gas emissions." These

technologies are recognized as emission reduction

certified by the ETS. A GHG technology government

agency converts KoreanOfset Credit (KOC) into

Korean Credit Unit (KCU) and trades at

approximately $19/tCO2e. technology was proposed

to reduce greenhouse gas emissions and a database

was created to convert them into economic value. The

raw data built by the method was used to develop a

model for selecting technologies to reduce

greenhouse gas emissions using machine learning.

3 RESULTS AND DISCUSSION

The methods used to calculate emission reductions

from GHG technology include a direct calculation

using variables and another one using statistical data

(Ministry of Environment, 2017). Based on the

established baseline greenhouse gas abatement

technology unit data, we selected the modeling

methods used for supervised learning in machine

learning and performed the analysis. Although there

are many simulation methods for this type of learning,

GBRT, SVM, and DNN methods were chosen in our

study, which were used to develop predictive models

in previous studies in Korea and abroad. In order to

MMTGE 2022 - I International Conference "Methods, models, technologies for sustainable development: agroclimatic projects and carbon

neutrality", Kadyrov Chechen State University Chechen Republic, Grozny, st. Sher

166

identify the phenomena represented by the collected

data and any problems, we performed exploratory

data analysis (EDA) and data pre-processing. In order

to calculate the certified emission reduction for a

specific GHG emission reduction technology, it is

first necessary to determine the amount of energy

produced or reduced by this technology. In our study,

we proposed a direct assessment of energy production

and reduction in accordance with the application of

greenhouse gas emission reduction technology. The

results were used as input data for the evaluation

database. Solar energy refers to a method of

generating electricity that directly converts solar

energy into electrical energy using the photovoltaic

effect. The annual electricity production from solar

energy can be calculated using the capacity, the

number of installed systems, the electricity utilization

rate and the operation time factors (Peng, 2013).

Solar thermal energy systems harvest solar energy

to heat or preheat water and are often used as the

energy source for water heating. The annual amount

of energy produced by solar thermal collectors can be

calculated by taking into account the installed area of

collectors, the number of households with collectors,

the total reduction from the energy source and the

number of working days. Geothermal energy refers to

the energy of the earth, including hot water and rocks

located deep underground. It is used as an energy

source for cooling and heating buildings. Calculation

of energy production for solar energy, solar thermal

energy and geothermal energy was made by applying

the operating time and electricity utilization factor to

the installed capacity of these sources. Equipment

efficiency among daily storage application rate

calculations was calculated using the statistical

estimates provided in the Greenhouse Gas Emission

Reduction Plan Implementation Guide of the

Ministry of Environment and Renewable Energy of

the National Administrative Urban Construction

Agency. When energy production from GHG

emission reduction technology is not monitored, the

daily storage application factor was calculated taking

into account the load factor, average daily operation

time and geothermal efficiency factor (GFC). When

electricity is used as an energy source for heating and

cooling energy of a building, the geothermal

efficiency factor (GFC) is calculated taking into

account the load factor, the average daily operation

time and the geothermal heating efficiency factor, and

a unit conversion factor of 860 (Mcal/MWh) was

applied (Porfiriev, 2010). The cooling energy GFC

was calculated taking into account the geothermal

heating efficiency factor and the geothermal cooling

efficiency factor. For SFC, the daily amount of solar

heat storage, it was calculated by applying daily solar

radiation, heat collector efficiency and the same unit

conversion factor as GFC.

In order to evaluate high-efficiency equipment in

projects certified by ETS as emission reduction for

government agencies, we separated external

greenhouse gas emission reduction projects that used

high-efficiency equipment, high-efficiency lighting,

LED street lighting installation, and green roofs

(Nikoláeva, 2018). For high-efficiency equipment

such as green rooftops, high-efficiency lighting, LED

street lighting, electric vehicles and natural gas

vehicles, the amount of energy saved was calculated

by comparing the energy consumption before and

after the project. The energy consumption of high-

efficiency lighting and LED street lamps was

calculated using electricity consumption, the number

of luminaires installed and the time the lights were

turned off. Based on the report on the supply and use

of lighting equipment prepared by the Ministry of

Commerce, Industry and Energy and the Korea

Energy Agency, high-efficiency lighting is 5.3 hours,

and the installation of LED street lighting is 10 hours

without light (Porfiriev, 2010). The GHG emission

reductions were calculated by applying the GHG

emission factors for each energy source to the energy

production results and reductions. In order to

establish a standard environmental assessment

database for greenhouse gas emission reduction

projects, in our study, the amount of energy produced

and reduced by greenhouse gas emission reduction

technology was converted to greenhouse gas

emission reduction when developing the assessment

database with greenhouse gas emissions as base unit.

4 CONCLUSIONS

After studying 1199 technology projects to reduce

greenhouse gas emissions, a method was proposed for

estimating the amount of energy reduced and

produced, and a certified greenhouse gas emission

reduction (KOC) method was proposed. Using 1199

GHG technology projects, SAM was created, which

is a GHG technology assessment database. To

consider energy consumption patterns and

environmental conditions in a building, SAM was

created to evaluate the effect of reducing greenhouse

gas emissions across different uses and energy

sources. These include heating, cooling, lighting,

ventilation and water heating, as well as energy

sources such as electricity, city gas and heat. Based

on the original SAM data, machine learning methods

(GBRT, SVM and DNN) were used to develop

Machine Learning Decision Support Model for Greenhouse Gas Reduction Technology Application

167

GRTM, a model that supports decision making for

greenhouse gas reduction technologies.

REFERENCES

Korea Energy Agency. 2018 Korea Energy Agency

Handbook.

Ministry of Environment. A Report on Korean-Sustainable

Development Goals(K-SDGs). 2019.

Ministry of Environment. Act on The Allocation and

Trading of Greenhouse-Gas Emission Permits,

Enforcement. 2017.

Ministry of Environment. Guidelines for Feasibility

Assessment and Certiﬁcation of External Projects.

2018.

Chan, M. S., 2017. Selection Model for Optimal Target

Building for Green Remodeling Using Data-Mining

Techniques: Focused on Multi-Family Housing

Complex. Yonsei University, Seoul, Korea.

Jang, B. G., Lee, S. J., 2018. Prediction and optimization

modeling using machine learning algorithm for

renewable energy. Inst. Electr. Inf. Eng. 6. pp. 1477-

1480.

Nakano, R., Zusman, E., Nugroho, S., Kaswanto, R., Ariﬁn,

N., Munandar, A., Ariﬁn, H. S., Muchtar, M., Gomi, K.,

Fujita, T., 2018. Determinants of energy savings in

Indonesia: The case of LED lighting in Bogor. Sustain.

Cities Soc. 42. pp. 184-193.

Singh, D., Basu, C., Meinhardt-Wollweber, M., Roth, B.

W., 2015. LEDs for energy eﬃcient greenhouse

lighting. Renew. Sustain. Energy Rev. 49. pp. 139-147.

Bottou, L., Curtis, F., Nocedal, J. Optimization methods for

large-scale machine learning. SIAM Rev. 2(60). pp.

223-311.

Jordan, M. I., Mitchell, T. M. Machine learning: Trends,

perspectives, and prospects. Science. 201, 349. pp. 255-

260.

Korea Energy Agency. Guidance for Clean Development

Mechanism. 2018.

Ministry of Environment. Guidance for Greenhouse Gas

Energy Target Management System. 2017.

Peng, J., Lu, L., Yang, H., 2013. Review on life cycle

assessment of energy payback and greenhouse gas

emission of solar photovoltaic systems. Renew. Sustain.

Energy Rev. 19. pp. 255-274.

Porfiriev, B., 2010. Climate change: risks or development

factors Russia in global politics.

http://www.globalaffairs.ru/number/Atmosfera-i-

ekonomika-14886.

Nikoláeva, L. B., 2018. Latin American economy in the face

of climate changes. New priorities.

MMTGE 2022 - I International Conference "Methods, models, technologies for sustainable development: agroclimatic projects and carbon

neutrality", Kadyrov Chechen State University Chechen Republic, Grozny, st. Sher

168