Sistem Prediksi Kondisi Mesin Jet Loader Menggunakan Metode Regresi Linear di PT Toyota Boshoku Indonesia

Ginting, Aqilah (2026) Sistem Prediksi Kondisi Mesin Jet Loader Menggunakan Metode Regresi Linear di PT Toyota Boshoku Indonesia. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040221005-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

PT Toyota Boshoku Indonesia menggunakan Jet Loader pada lini injection molding untuk menyalurkan material plastik dari dryer tank menuju hopper. Sistem perawatan yang masih bersifat reaktif menyebabkan perubahan kondisi mesin sebelum terjadi gangguan belum dapat diketahui secara kuantitatif dan berkelanjutan. Selain itu, teknisi belum memperoleh informasi mengenai perkiraan waktu operasi tersisa sebelum Jet Loader mencapai kondisi kritis, sehingga penentuan waktu perawatan preventif masih bergantung pada pemeriksaan langsung dan pengalaman teknisi. Penelitian ini menerapkan sistem pemantauan kondisi Jet Loader berdasarkan empat parameter, yaitu tekanan vakum, arus motor, suhu pompa, dan getaran pompa. Keempat parameter tersebut diolah menjadi Health Index sebagai indikator kondisi mesin. Tren perubahan Health Index kemudian dianalisis menggunakan regresi linear untuk memperkirakan Remaining Useful Life. Hasil pemantauan dan estimasi ditampilkan melalui dashboard serta disampaikan melalui notifikasi WhatsApp sebagai informasi pendukung bagi teknisi. Hasil pengamatan periode 1–19 Desember 2025 menunjukkan bahwa nilai Health Index menurun dari 0,75 menjadi 0,32. Penurunan tersebut sejalan dengan melemahnya tekanan vakum dari −0,23 bar menjadi −0,21 bar, meningkatnya arus motor dari 5,81 A menjadi 6,08 A, suhu pompa dari 60,63°C menjadi 65,20°C, serta getaran pompa dari 1,39 g menjadi 2,13 g. Model regresi menghasilkan slope sebesar −0,02 per hari. Pada skenario validasi terbaik, diperoleh RMSE sebesar 0,03, MAPE sebesar 7,48%, dan R² sebesar 0,80. Kondisi kritis diprediksi terjadi pada 19 Desember 2025, sedangkan data aktual mencapai batas kritis pada 18 Desember 2025. Validasi pengguna memperoleh nilai 87,67%. Hasil tersebut menunjukkan bahwa sistem mampu memberikan informasi mengenai penurunan kondisi dan estimasi waktu menuju kondisi kritis sebagai pendukung pelaksanaan perawatan preventif Jet Loader.
===================================================================================================================================
PT Toyota Boshoku Indonesia uses a Jet Loader on its injection molding line to transfer plastic material from the dryer tank to the hopper. The current reactive maintenance approach makes changes in machine condition difficult to identify quantitatively and continuously before a failure occurs. In addition, maintenance technicians do not yet have information regarding the estimated remaining operating time before the Jet Loader reaches a critical condition, so preventive maintenance scheduling still relies on direct inspection and technician experience. This study applies a Jet Loader condition monitoring system based on four parameters: vacuum pressure, motor current, pump temperature, and pump vibration. These parameters are processed into a Health Index as an indicator of machine condition. The trend of the Health Index is then analyzed using linear regression to estimate the Remaining Useful Life. The monitoring and estimation results are presented through a dashboard and delivered via WhatsApp notifications as supporting information for maintenance technicians. The observation results from 1 to 19 December 2025 show that the Health Index decreased from 0.75 to 0.32. This decrease was accompanied by a weakening of vacuum pressure from −0.23 bar to −0.21 bar, an increase in motor current from 5.81 A to 6.08 A, an increase in pump temperature from 60.63°C to 65.20°C, and an increase in pump vibration from 1.39 g to 2.13 g. The regression model produced a slope of −0.02 per day. In the best validation scenario, the model achieved an RMSE of 0.03, a MAPE of 7.48%, and an R² of 0.80. The critical condition was predicted to occur on 19 December 2025, while the actual data reached the critical threshold on 18 December 2025. User validation obtained a score of 87.67%. These results indicate that the system can provide information on Jet Loader condition degradation and estimate the remaining operating time before reaching the critical condition to support preventive maintenance.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Jet Loader, Health index, Regresi Linear, Remaining Useful Life, Perawatan Prediktif. Keywords: Jet Loader, Health index, Linear Regression, Remaining Useful Life, Predictive Maintenance.
Subjects:	T Technology > TJ Mechanical engineering and machinery > TJ217.6 Predictive Control
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Aqilah Tsabitah Ginting
Date Deposited:	10 Aug 2026 01:33
Last Modified:	10 Aug 2026 01:33
URI:	http://repository.its.ac.id/id/eprint/144245

Actions (login required)

View Item