Shamara, Nadya Mariela (2023) Sistem Klasifikasi Batu Bijih Emas Berdasarkan Distribusi Ukuran Partikel Menggunakan Convolutional Neural Network Untuk Untuk Efisiensi Proses Grinding Pada Semi-Autogenous Mill. Other thesis, Institut Teknologi Sepuluh Nopember Surabaya.
Text
4206100516-Undregraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (8MB) | Request a copy |
Abstract
Dalam proses grinding, komposisi atau distribusi batu bijih emas yang akan masuk ke Semi-Autogenous Mill (SAG-Mill) sering kali tidak seimbang komposisinya. Seringkali mining engineer salah memperkirakan ukuran batu sehingga proses grinding tersebut tidak berjalan lancar dan membutuhkan waktu yang lama. Pada proses grinding, batuan besar akan menghancurkan batuan medium, dan batuan medium akan menghancurkan batuan kecil sehingga distribusi ukuran batu bijih emas merupakan dasar penting untuk mengevaluasi efek dari penghancuran batu bijih emas yang dilakukan oleh perusahaan tambang, dan juga merupakan indeks utama untuk kontrol mineral yang optimal dari pengolahan emas. Sehingga, diperlukan suatu sistem yang dapat mengklasifikasikan batu bijih emas dan dipilah melalui ukuran agar bisa dipantau distribusinya. Salah satu metode yang dapat diterapkan dalam mengklasifikasikan suatu objek gambar batu bijih emas untuk berdasarkan distribusi ukuran diterapkan metode Convolutional Neural Network (CNN), data diuji dengan berbagai arsitektur CNN seperti ResNet50, VGG16, Inception V3 dan Xception menggunakan Transfer Learning, dari hasil pengujian yang didapatkan VGG16 menjadi arsitektur yang memiliki tingkat akurasi terbaik sekitar 92% dibandingkan arsitektur lainnya. Dalam penelitian ini juga mengunjukkan bahwa pengaruh sistem klasifikasi batu bijih menggunakan arsitektur VGG16 dalam menentukan distribusi ukuran partikel mempengaruhi peningkatan efisiensi proses grinding pada SAG-Mill
======================================================================================================================================
In the grinding process, the composition or distribution of gold ore that will enter the Semi-Autogenous Mill (SAG-Mill) is often not balanced in composition. Often mining engineers incorrectly estimate the size of the stone so that the grinding process does not run smoothly and takes a long time. In the grinding process, large rocks will crush medium rocks, and medium rocks will crush small rocks so that the size distribution of gold ore rocks is an important basis for evaluating the effects of gold ore rock crushing carried out by mining companies, and is also the main index for mineral control. Optimization of gold processing. So, we need a system that can classify gold ore rock and sort it by size so that its distribution can be monitored. One method that can be applied in classifying a gold ore image object based on size distribution is the Convolutional Neural Network (CNN) method, the data is tested with various CNN architectures such as ResNet50, VGG16, Inception V3 and Xception using Transfer Learning, from the test results It was found that VGG16 is the architecture that has the best accuracy rate of around 92% compared to other architectures. This research also shows that the influence of the ore rock classification system using the VGG16 architecture in determining particle size distribution influences the increase in efficiency of the grinding process at the SAG-Mill
Item Type: | Thesis (Other) |
---|---|
Uncontrolled Keywords: | Deep learning, Convolutional Neural Network, Transfer Learning, SAG-Mill, Ore Mining, Image analysis, Minerals processing |
Subjects: | T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK3070 Automatic control T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK4055 Electric motor T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.546 Computer algorithms |
Divisions: | Faculty of Vocational > 36304-Automation Electronic Engineering |
Depositing User: | NADYA MARIELA SHAMARA |
Date Deposited: | 03 Jun 2024 00:48 |
Last Modified: | 03 Jun 2024 00:48 |
URI: | http://repository.its.ac.id/id/eprint/108014 |
Actions (login required)
View Item |