Prediksi Kadar Sulfur Electric Furnace Matte Mengunakan Machine Learning

Winoto, Gatot (2022) Prediksi Kadar Sulfur Electric Furnace Matte Mengunakan Machine Learning. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6010201025-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2024. Download (2MB) | Request a copy

Abstract

Kadar sulfur pada electric furnace matte adalah salah satu variabel kunci dalam pengendalian proses pada industri pembuatan nickel matte di PT Vale Indonesia (PTVI). Saat ini, pemenuhan terhadap standar spesifikasi kadar sulfur dalam electric furnace matte masih cukup rendah akibat variasi proses dan keterbatasan kontrol pengendalian proses. Kadar sulfur di electric furnace matte akan sangat tergantung dari penambahan sulfur dan kondisi operasi di tanur pereduksi dan tanur peleburan. Dalam penelitian ini, sebuah pendekatan machine learning digunakan untuk membuat model prediksi kadar sulfur dalam electric furnace matte berdasarkan variabel input yang dipilih. Model regresi linier dan regresi support vector dibuat, dievaluasi dan dibandingkan. Model regresi linier menunjukkan koefisien korelasi 0,5843 antara data uji dan data prediksi model dengan mean square error (MSE) 0,4207. Model regresi support vector, sebuah model non-liniear, dibuat dengan pendekatan yang sama. Dari hasil pemodelan diperoleh bahwa model regresi support vector dapat meningkatkan koefisien korelasi menjadi 0,9408 dan mengurangi MSE menjadi 0,0762. Model regresi support vector berhasil memprediksi kadar sulfur pada electric furnace matte dengan kinerja yang bagus dan bisa digunakan dalam pengendalian kadar sulfur di electric furnace matte.
=================================================================================================================================
Sulfur content in electric furnace matte is one of the key parameter in nickel matte process control at PT Vale Indonesia (PTVI), a nickel matte smelter in Indonesia. Currently the compliance of sulfur content in electric furnace matte to standard specification is relatively low due to process variability and control limitation. The sulfur content in electric furnace matte depends on sulfur addition and other equipment operating conditions in kiln and furnace. In this reserach, a machine learning approach is used to predict sulfur content in electric furnace matte based on selected input variables. Linear and support vector regression models are built on the training data and used to predict sulfur content on testing data. The performance of each modelsis evaluated and compared. The linear model shows a 0,5843 coefficient correlation between test data and prediction, with a mean square error (MSE) 0.4207. The support vector regression (SVR), a nonlinier model, is built with the same approach. SVR model improve the correlation coefficient to 0,9408 and reduce the MSE to 0,0762. Support vector regression model can provide better performance and can be used in sulfur process control in electric furnace matte.

Item Type:	Thesis (Masters)
Additional Information:	RTI 006.312 Win p-1 2022
Uncontrolled Keywords:	machine learning, nickel matte, linier regression, support vector regression
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science)
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26101-(S2) Master Thesis
Depositing User:	Anis Wulandari
Date Deposited:	12 Jun 2024 02:53
Last Modified:	12 Jun 2024 02:53
URI:	http://repository.its.ac.id/id/eprint/108035

Actions (login required)

View Item