Prediksi Kadar Abu dan Kadar Air Batubara Menggunakan Spektra FTIR-ATR

Wahyuningtyas, Antin (2023) Prediksi Kadar Abu dan Kadar Air Batubara Menggunakan Spektra FTIR-ATR. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 01211940000079-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (3MB) | Request a copy

Abstract

Penelitian prediksi kadar abu dan kadar air pada 20 jenis sampel uji batubara menggunakan spektra FTIR-ATR telah berhasil dilakukan. Sampel uji diperoleh dari Laboratorium Energi dan Lingkungan ITS. Analisis kadar abu dan kadar air pada sampel uji batubara diuji dengan metode SNI. Hasil dari pengujian menggunakan metode SNI diperoleh rata-rata kadar air sebesar 6,40% dengan nilai simpangan baku sebesar 2,51%. Nilai minimum dari data kadar air 20 sampel uji batubara yaitu sebesar 2,73% dengan nilai kuartil 1 (Q1) sebesar 4,39%, kuartil 2 (Q2) sebesar 6,03%, (Q3) sebesar 8,14% dan nilai maksimum sebesar 10,77%. Pada hasil pengujian kadar abu menggunakan metode SNI, nilai rata-rata kadar abu yaitu sebesar 18,72% dengan nilai simpangan baku sebesar 7,73%. Nilai minimum dari data kadar abu 20 sampel uji batubara yaitu sebesar 4,85% dengan nilai kuarti 1 (Q1) sebesar 13,49%, kuartil 2 (Q2) sebesar 17,47%, kuartil 3 (Q3) sebesar 25,06% dan nilai maksimum sebesar 32,56%. Selanjutnya, hasil analisis kadar abu dan kadar air dari sampel uji batubara menggunakan metode SNI dibandingkan dengan hasil prediksi kadar abu dan kadar air menggunakan data spektra FTIR-ATR dari 20 sampel uji batubara. Prediksi kadar abu dan kadar air menggunakan spektra FTIR-ATR pada 20 jenis sampel uji batubara dilakukan dengan memecah data dengan machine learning dimana 75% data spektra sebagai data untuk kalibrasi (training set) dan 25% data spektra lainnya untuk memvalidasi model regresi untuk memprediksi kadar air dan kadar abu. Hasil prediksi kadar abu menggunakan model regresi Lasso Lars CV dapat menjelaskan varian kadar abu untuk data uji (test data) dengan menunjukkan nilai R2 sebesar 0,96 dan adjusted R2 sebesar 1,00. Hasil prediksi kadar air menggunakan model regresi Lasso Lars CV dapat menjelaskan varian kadar air untuk data uji (test data) dengan menunjukkan nilai R2 sebesar 0,994 dan adjusted R2 sebesar 1,00.
===============================================================================================================================
Twenty coal samples were analyzed for ash and water content predictions using FTIR-ATR spectra. These samples were sourced from the Energy and Environment Laboratory at ITS. The SNI method was applied to analyze the ash and water content in these coal test samples. The SNI method testing yielded interesting results. The average water content in the coal test samples was 6,40%, with a standard deviation of 2,51%. The range of water content varied, with the lowest value recorded at 2,73%. The distribution across quartiles was as follows: the first quartile (Q1) at 4,39%, the second quartile (Q2) at 6,03%, the third quartile (Q3) at 8,14%, and the maximum value at 10,77%. Regarding the ash content testing using the SNI method, the average ash content was 18,72%, with a standard deviation of 7,73%. The minimum ash content value among the 20 coal test samples was 4,85%. The quartile distribution for ash content was as follows: the first quartile (Q1) at 13,49%, the second quartile (Q2) at 17,47%, the third quartile (Q3) at 25,06%, and the maximum value at 32,56%. To validate the FTIR-ATR spectra data, the predicted results of ash and water content from the 20 coal test samples were compared with the analyses using the SNI method. The predictions were obtained using a machine learning approach, with 75% of the spectra data used for calibration (training set) and the remaining 25% for validating the regression model's accuracy in predicting water and ash content. Impressively, the ash content prediction using the Lasso Lars CV regression model showed a remarkable ability to explain the variance in the test data, with an R2 value of 0,96 and an adjusted R2 value of 1,00. Similarly, the water content prediction using the Lasso Lars CV regression model exhibited high accuracy in explaining the variance of water content for the test data, with an R2 value of 0,994 and an adjusted R2 value of 1,00

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Batubara, Kadar Air, Kadar Abu, FTIR-ATR,Coal, Moisture Content, Ash Content, ATR-FTIR
Subjects:	Q Science > QD Chemistry > QD117.S64 Spectrophotometry Q Science > QD Chemistry > QD75.2 Chemistry, Analytic
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Chemistry > 47201-(S1) Undergraduate Thesis
Depositing User:	Antin Wahyuningtyas
Date Deposited:	10 Aug 2023 14:06
Last Modified:	10 Aug 2023 14:06
URI:	http://repository.its.ac.id/id/eprint/104379

Actions (login required)

View Item