Pemodelan Prediksi Umpan Pulverized Coal Pada Sistem Rotary Kiln Pabrik Semen Berdasarkan Parameter Operasi & Kimia Berbasis Machine Learning

Maliki, Riduwan (2024) Pemodelan Prediksi Umpan Pulverized Coal Pada Sistem Rotary Kiln Pabrik Semen Berdasarkan Parameter Operasi & Kimia Berbasis Machine Learning. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6032222046-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2027. Download (4MB) | Request a copy

Abstract

Industri semen di Indonesia mengalami persaingan ketat dan tuntutan efisiensi operasional akibat kenaikan biaya energi. Konsumsi pulverized coal (PC) merupakan salah satu komponen biaya terbesar dalam produksi semen, dengan potensi optimasi yang signifikan. Penelitian ini bertujuan untuk mengembangkan model prediksi umpan pulverized coal (PC) pada sistem rotary kiln di pabrik semen menggunakan algoritma machine learning. Algoritma yang digunakan untuk pemodelan pada penelitian ini adalah Support Vector Machine (SVM), Neural Network , dan Gradient Boosting. Hasil pemodelan ini diharapkan dapat membantu operator Control Center Room (CCR) dalam mengoptimalkan konsumsi PC dan meningkatkan efisiensi produksi semen. Penelitian ini berfokus pada pemodelan prediksi umpan PC berbasis machine learning pada rotary kiln dengan memanfaatkan data parameter operasi, dan sifat kimia bahan baku dan produk dengan menggunakan algoritma machine learning serta menentukan faktor apa yang paling signifikan dalam mempengaruhi nilai umpan PC pada rotary kiln. Hasil penelitian menunjukkan bahwa algoritma Gradient Boosting memberikan akurasi terbaik dengan skor dari data training MSE = 0.101, RMSE = 0.317, MAE = 0.185, dan R2 = 0.948 dan skor dari data test MSE = 0.075, RMSE = 0.274, MAE = 0.169, dan R2 = 0.979. Model Gradient Boosting dan Neural Network mampu memberikan nilai Pulverized kiln secara rata-rata lebih rendah dibandingkan dengan pengaturan manual oleh operator, yaitu sebesar 0.2% lebih rendah untuk Gradient Boosting dan 0.14% lebih rendah untuk Neural Network. Implikasi dari hasil ini adalah potensi penghematan penggunaan batubara dan pengurangan emisi CO2 dalam proses produksi semen.
=================================================================================================================================
The Indonesian cement industry faces intense competition and demands for operational efficiency due to rising energy costs. Pulverized coal (PC) consumption is one of the largest cost components in cement production, with significant optimization potential. This research aims to develop a prediction model for pulverized coal feed in a rotary kiln system of a cement plant using machine learning algorithms. The algorithms used for modeling in this study are Support Vector Machine (SVM), Neural Network, and Gradient Boosting. The results of this modeling are expected to assist Control Center Room (CCR) operators in optimizing PC consumption and improving cement production efficiency. This research focuses on developing a machine learning-based prediction model for pulverized coal (PC) feed in a rotary kiln system. The model utilizes operational parameter data, as well as the chemical properties of raw materials and products. The study is to identify the most significant factors influencing PC feed values in the rotary kiln. The results demonstrate that the Gradient Boosting algorithm achieved the highest accuracy, with training scores of MSE = 0.101, RMSE = 0.317, MAE = 0.185, and R² = 0.948, and testing scores of MSE = 0.075, RMSE = 0.274, MAE = 0.169, and R² = 0.979. Both the Gradient Boosting and Neural Network models were able to predict lower Pulverized Kiln feed rates compared to manual operator adjustments, with Gradient Boosting achieving 0.2% reduction and Neural Network 0.14% reduction. These findings imply the potential for significant coal consumption savings and reduced CO2 emissions in the cement production.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	rotary kiln, cement, coal, energy, machine learning, Support Vector Machine (SVM), Gradient Boosting, Neural Network, semen, batubara, energi, mesin pembelajaran
Subjects:	Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA9.58 Algorithms
Divisions:	Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT)
Depositing User:	Riduwan Maliki
Date Deposited:	02 Dec 2024 04:11
Last Modified:	02 Dec 2024 04:11
URI:	http://repository.its.ac.id/id/eprint/115868

Actions (login required)

View Item