Sistem Penentuan Jadwal Perawatan Mesin Truckloader Berdasarkan Nilai Keandalan Menggunakan Distribusi Weibull

Amujianto, Ayyub (2026) Sistem Penentuan Jadwal Perawatan Mesin Truckloader Berdasarkan Nilai Keandalan Menggunakan Distribusi Weibull. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040221041-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (11MB) | Request a copy

Abstract

Mesin truckloader pada lini pengantongan pupuk berfungsi untuk memindahkan pupuk ke gudang maupun langsung ke bak truk. Kerusakan mesin truckloader pada komponen conveyor belt menjadi penyebab gangguan produksi yang paling sering terjadi, dengan tujuh kejadian kerusakan per tahun berdasarkan data historis. Ketiadaan jadwal perawatan yang terencana membuat tim pemeliharaan seringkali hanya bereaksi setelah kerusakan terjadi, sehingga menimbulkan downtime yang tidak terencana dan berpotensi mengganggu target produksi. Penelitian ini bertujuan mengembangkan sistem penentuan jadwal perawatan mesin truckloader berdasarkan nilai keandalan yang memprediksi tanggal kerusakan berikutnya, sehingga perawatan dapat dilakukan sebelum mesin mengalami kegagalan. Pengembangan sistem dilakukan menggunakan data operasional mesin periode Januari 2024 hingga Juli 2025. Riwayat kerusakan conveyor belt yang terjadi selama periode tersebut dimodelkan menggunakan distribusi Weibull untuk menentukan waktu mesin ambang batas nilai keandalan. Selanjutnya, simulasi Monte Carlo digunakan untuk memprediksi jadwal perawatan mesin berikutnya dengan mempertimbangkan variasi pola operasi aktual mesin pada setiap shift, sehingga prediksi yang dihasilkan lebih adaptif dibandingkan pendekatan deterministik sederhana. Pengujian sistem dilakukan dengan melatih model menggunakan tujuh data interval kerusakan terdahulu, kemudian hasilnya digunakan untuk memprediksi kerusakan pada interval 8–9 yang tidak pernah digunakan dalam proses pemodelan. Hasil pengujian menunjukkan bahwa estimasi persentil P1, yang digunakan sebagai acuan utama rekomendasi jadwal perawatan berdasarkan pendekatan paling konservatif dalam pemeliharaan berbasis keandalan, menghasilkan waktu prediksi yang sesuai dengan waktu kerusakan aktual sehingga diperoleh error sebesar 0 hari. Sistem kemudian diimplementasikan pada dashboard monitoring yang menampilkan rekomendasi jadwal perawatan berdasarkan data running hours mesin, sehingga dapat digunakan oleh operator maupun teknisi sebagai alat bantu perencanaan perawatan.
===================================================================================================================================
The truckloader machine on the fertilizer bagging line functions to transport fertilizer to the warehouse or directly onto trucks. Failures of the truckloader, particularly in the conveyor belt component, are the most frequent cause of production disruptions, with seven failure events per year based on historical data. The absence of a planned maintenance schedule causes the maintenance team to rely on corrective actions after failures occur, resulting in unplanned downtime that may disrupt production targets. This study aims to develop a reliability-based maintenance scheduling system for the truckloader machine by predicting the next failure time, enabling maintenance to be performed before machine failure occurs. The system was developed using the machine's operational data from January 2024 to July 2025. The conveyor belt failure history during this period was modeled using the Weibull distribution to determine the critical reliability threshold time. Subsequently, Monte Carlo simulation was employed to predict the maintenance schedule by considering variations in the machine's actual operating patterns across different shifts, resulting in a more adaptive prediction than a simple deterministic approach. The system was evaluated by training the model using seven previous failure intervals and then predicting failures for intervals 8–9, which were not included in the model development process. The evaluation results show that the P1 percentile estimate, used as the primary reference for maintenance scheduling based on the most conservative reliability centered maintenance planning approach, produced a prediction that matched the actual failure time with a prediction error of 0 days. The system was implemented as a monitoring dashboard that displays maintenance schedule predictions based on the machine's running hours, providing operators and maintenance technicians with a practical tool for maintenance planning.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Keandalan mesin, Distribusi Weibull, Prediksi jadwal perawatan mesin, Machine reliability, Weibull distribution, Maintenance schedule prediction.
Subjects:	Q Science > QA Mathematics > QA273.6 Weibull distribution. Logistic distribution. T Technology > TJ Mechanical engineering and machinery > TJ174 Maintenance and repair of machinery
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Ayyub Amujianto
Date Deposited:	05 Aug 2026 02:32
Last Modified:	05 Aug 2026 02:32
URI:	http://repository.its.ac.id/id/eprint/143829

Actions (login required)

View Item