RANCANG BANGUN KERANGKA KERJA MULTI-MODEL MACHINE LEARNING UNTUK SISTEM PREDICTIVE MAINTENANCE DENGAN ANTARMUKA BERBASIS WEBSITE

Andreas Malem Sebayang, - and Iwan Kustiawan, - (2025) RANCANG BANGUN KERANGKA KERJA MULTI-MODEL MACHINE LEARNING UNTUK SISTEM PREDICTIVE MAINTENANCE DENGAN ANTARMUKA BERBASIS WEBSITE. S1 thesis, Universitas Pendidikan Indonesia.

Abstract

Industri 4.0 mendorong transformasi menuju ekosistem berbasis teknologi dan data, di mana pemeliharaan prediktif menjadi salah satu keunggulan utama untuk meminimalkan waktu henti dan biaya perawatan. Namun, integrasi berbagai model machine learning dengan fungsi berbeda ke dalam satu dashboard masih menjadi tantangan. Penelitian ini mengembangkan sistem pemeliharaan prediktif berbasis website, yang mengintegrasikan tiga model unggulan, seperti extreme gradient boosting (XGBoost) untuk peramalan data operasional, isolation forest (IForest) untuk deteksi anomali, dan random forest (RF) untuk prediksi kegagalan perangkat. Proses mencakup pengumpulan data sensor dan historis kegagalan, pembersihan data, pelatihan dan tuning model, evaluasi performa, dan pengembangan dashboard berbasis website. Hasil menunjukkan sistem mampu memberikan prediksi kondisi perangkat dengan akurasi tinggi dan kemudahan integrasi melalui dashboard, meningkatkan efisiensi dan keandalan pengelolaan peralatan industri. Industry 4.0 drives the transformation toward a technology and data-driven ecosystem, where predictive maintenance becomes a key advantage to minimize downtime and maintenance costs. However, integrating multiple machine learning models with different functions into a single dashboard remains a challenge. This study develops a web-based predictive maintenance system, which integrates three leading models: extreme gradient boosting (XGBoost) for operational data forecasting, isolation forest (IForest) for anomaly detection, and random forest (RF) for equipment failure prediction. The process includes collecting sensor and historical failure data, data cleaning, model training and tuning, performance evaluation, and the development of an interactive web-based dashboard. The results show that the system can provide highly accurate predictions of equipment conditions and easy integration through the dashboard, improving the efficiency and reliability of industrial equipment management.

	Text S_TE_2004093_Title.pdf Download (544kB)
	Text S_TE_2004093_Chapter1.pdf Download (248kB)
	Text S_TE_2004093_Chapter2.pdf Restricted to Staf Perpustakaan Download (526kB) | Request a copy
	Text S_TE_2004093_Chapter3.pdf Download (375kB)
	Text S_TE_2004093_Chapter4.pdf Restricted to Staf Perpustakaan Download (701kB) | Request a copy
	Text S_TE_2004093_Chapter5.pdf Download (226kB)
	Text S_TE_2004093_Appendix.pdf Restricted to Staf Perpustakaan Download (382kB) | Request a copy

Item Type:	Thesis (S1)
Additional Information:	https://scholar.google.com/citations?hl=en&user=KcstBjAAAAAJ ID SINTA Dosen Pembimbing: Iwan Kustiawan: 5996452
Uncontrolled Keywords:	Pemeliharaan prediktif, peramalan, deteksi anomali, prediksi kegagalan, Extreme Gradient Boosting (XGBoost), Isolation Forest (IForest), Random Forest (RF), Website. Predictive Maintenance, Forecasting, Anomaly Detection, Failure Prediction, Extreme Gradient Boosting (XGBoost), Isolation Forest (IForest), Random Forest (RF), Website.
Subjects:	L Education > L Education (General) T Technology > T Technology (General) T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions:	Fakultas Pendidikan Teknik dan Industri > Jurusan Pendidikan Teknik Elektro > Program Studi Teknik Tenaga Elektrik
Depositing User:	Andreas Malem Sebayang
Date Deposited:	09 Dec 2025 07:09
Last Modified:	09 Dec 2025 07:09
URI:	http://repository.upi.edu/id/eprint/145402

Actions (login required)

View Item