Pemodelan Prediksi Emisi Sulfur Dioksida (SO2), Nitrogen Oksida (NOx) dan Partikulat dari Pembangkit Listrik Tenaga Uap Berbahan Bakar Batubara dengan Machine Learning

Gaol, Rina Yani Lumban (2023) Pemodelan Prediksi Emisi Sulfur Dioksida (SO2), Nitrogen Oksida (NOx) dan Partikulat dari Pembangkit Listrik Tenaga Uap Berbahan Bakar Batubara dengan Machine Learning. Masters thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 6032211008-Master_Thesis.pdf] Text
6032211008-Master_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2025.

Download (1MB) | Request a copy

Abstract

Emisi dari sisa kegiatan pembakaran batubara pada operasional Pembangkit Listrik Tenaga Uap (PLTU) adalah Sulfur Dioksida (SO2), Nitrogen Oksida (NOx) dan partikulat yang merupakan salah satu sumber utama polusi udara yang berkontribusi terhadap masalah lingkungan dan kesehatan masyarakat. Adanya dampak negatif inilah, pemerintah Indonesia menetapkan peraturan untuk mengatur batasan emisi SO2 dan juga akan semakin memperketat nilai ambang batas emisi gas buang PLTU. Penelitian ini menggunakan tiga model prediksi machine learning, yaitu Gradient Boosting, Artificial Neural Network (ANN), dan Support Vector Regression (SVR). Data yang digunakan dalam penelitian ini berisi informasi tentang emisi SO2, NOx, dan partikulat dari PLTU yang menggunakan bahan bakar batubara berkapasitas 100 Mwh. Variabel independen terdiri dari 19 variabel diambil dari pemantauan soft sensor yang terpasang pada boiler serta teknologi pengendali emisi gas buang dan dipantau secara real time. Evaluasi performa model dilakukan dengan menggunakan metrik Root Mean Squared Error (RMSE), Coefficient of Determination (R2), dan Mean Absolute Error (MAE) Model prediksi terbaik yang didapatkan pada penelitian ini untuk memprediksi emisi SO2, NOx, dan partikulat dari pembangkit listrik tenaga uap berbahan bakar batubara adalah model Gradient Boosting. Model ini menghasilkan performa prediksi yang lebih baik dibandingkan dengan model ANN dan SVR, dilihat dari nilai RMSE dan MAE yang terkecil serta R2 paling optimal. Model prediksi dengan GB menghasilkan faktor yang mempengaruhi nilai emisi SO2.
===============================================================================================================================
The emission from the residual combustion activities in Coal-Fired Power Plants (CFPP) operation includes Sulfur Dioxide (SO2), Nitrogen Oxides (NOx), and particulate matter, which are major contributors to air pollution, leading to environmental and public health issues. In response to these negative impacts, Indonesian government has implemented regulations to control the emission limits and plans to further tighten the emission threshold for CFPP exhaust gases. This study employs three (3) machine learning prediction models, namely Gradient Boosting (GB), Artificial Neural Network (ANN) and Support Vector Regression (SVR). The dataset used in this study contains information on SO2, NOx, and particulate matter emissions from CFPP with capacity is 100 MWh. The independent variables consist of 19 variables taken from soft sensor monitoring installed on the boiler and emission control technology, which are continuously monitored in real-time. The performance evaluation of the models is done using metrics such as Root Mean Squared Error (RMSE), Coefficient of Determination (R2), and Mean Absolute Error (MAE). The best prediction model obtained in this study for predicting the emissions of SO2, NOx, and particulate matter from coal-fired power plants is Gradient Boosting. This model results in better prediction performance compared to the ANN and SVR models, as seen from the smallest RMSE and MAE values and the most optimal R square. GB prediction model shows that a factor influencing SO2 emissions is limestone frequency, NOx is furnace upper temperature, and particulate is main steam temperature.

Item Type: Thesis (Masters)
Uncontrolled Keywords: Gradient Boosting, ANN, SVR, RMSE
Subjects: Q Science > Q Science (General)
Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions: Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT)
Depositing User: Rina Yani Lumban Gaol
Date Deposited: 07 Sep 2023 06:56
Last Modified: 07 Sep 2023 06:56
URI: http://repository.its.ac.id/id/eprint/104108

Actions (login required)

View Item View Item