Prediksi Waktu Kegagalan Mesin Raw Mill Menggunakan Active Learning

Maulana, Thoriq Rizky (2023) Prediksi Waktu Kegagalan Mesin Raw Mill Menggunakan Active Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 06111940000084-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (4MB) | Request a copy

Abstract

Pemeliharaan mesin raw mill menjadi sangat penting dikarenakan mesin utama pertama dalam produksi semen. Untuk mendapatkan pemeliharaan mesin yang optimal, predictive maintenance perlu dilakukan untuk memprediksi kapan mesin dinyatakan gagal bekerja. Dengan bantuan machine learning, prediksi waktu kegagalan mesin dapat dilakukan secara otomatis. Penelitian Tugas Akhir ini bertujuan untuk memprediksi kegagalan mesin raw mill menggunakan pendekatan active learning. Penelitian terkait menggunakan pendekatan active learning berdasarkan ambang batas terhadap nilai loss dari data latih. Penentuan ambang batas sangat sulit dilakukan karena berbeda untuk setiap kejadian kegagalan. Oleh karena itu, peran operator mesin masih perlu dilakukan untuk mengurangi kesalahan prediksi kegagalan mesin. Penelitian Tugas Akhir ini mengusulkan sebuah model yang menggunakan operator mesin pada setiap prediksi kegagalan mesin dari metode active learning. Dalam eksperimen, model yang diusulkan diuji pada dataset sensor yang terinstal pada mesin raw mill pada salah satu pabrik semen di Indonesia. Hasil dari eksperimen akan menunjukkan dan memberikan prediksi kegagalan mesin raw mill dengan baik.
=================================================================================================================================
Maintenance on raw mill engine becomes crucial as it is the first major engine in sand-cement production. To have the optimal maintenance, we need to perform predictive maintenance to predict when the engine is broken. From machine learning, we can predict the failure time of engine automatically. This study aims to predict the failure time of raw mill engine using active learning model. Related studies utilized active learning model based on threshold that formulated from training loss. The formulation can be very difficult to fit for every failure conditions. Hence, the engine operator (human) is still needed to reduce the error of the failure time of engine. This study proposes a model by cooperating with engine operator in each engine failure occurred via the active learning method. In the experiments, the proposed model was tested on sensor data that embedded in raw mill engine in one of sand-cement factory in Indonesia. The results of the experiment can show and give the failure time prediction with good accuracy.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Mesin Raw Mill, Predictive Maintenance, kerusakan, Model Active Learning.
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Q Science > QA Mathematics
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Thoriq Rizky Maulana
Date Deposited:	26 Sep 2023 02:07
Last Modified:	26 Sep 2023 02:07
URI:	http://repository.its.ac.id/id/eprint/102731

Actions (login required)

View Item