Septian, Wayan Indra (2026) Sistem Prediksi Kegagalan Pompa Axial Piston Menggunakan Integrasi K-Means Dan Analisis Weibull Pada Line Polyurethane Pt Toyota Boshoku Indonesia. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 2040221102-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
2040221102-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy |
Abstract
PT Toyota Boshoku Indonesia (TBINA) menghadapi kendala gangguan berulang pada pompa axial piston di line polyurethane yang memicu downtime produksi akibat keterbatasan strategi pemeliharaan berbasis waktu. Penelitian ini bertujuan mengembangkan sistem predictive maintenance berbasis vibrational monitoring untuk mendukung implementasi Condition-Based Maintenance (CBM). Arsitektur sistem dibangun menggunakan sensor ADXL335 untuk akuisisi data real-time, ESP32 sebagai media transmisi, Raspberry Pi sebagai pusat pemrosesan data, serta Node-RED sebagai antarmuka pemantauan. Pengolahan data getaran meliputi tahap preprocessing, ekstraksi fitur statistik domain waktu, feature engineering, pembentukan Health Index, serta klasterisasi kondisi operasi menggunakan algoritma Gaussian Mixture Model (GMM) yang menghasilkan Silhouette Score sebesar 0,5948 untuk tiga kondisi utama. Selanjutnya, analisis reliabilitas berbasis distribusi Weibull pada interval precursor event menghasilkan parameter β = 0,9865, η = 64,6112 jam, dan Mean Time Between Precursor (MTBP) = 64,99 jam. Hasil validasi sistem menunjukkan performa analitis yang tinggi dengan tingkat Accuracy 98,54%, Recall 100,00%, Precision 50,00%, F1-score 66,67%, serta Average Lead Time selama 69,94 jam (≈ 3 hari) sebelum eksekusi pemeliharaan. Sistem ini terbukti andal memberikan peringatan dini guna mendukung penyusunan strategi CBM secara efektif di industri.
======================================================================================================================================
PT Toyota Boshoku Indonesia (TBINA) faces recurrent failure issues with axial piston pumps on its polyurethane production line, which trigger production downtime and disrupt manufacturing continuity. Traditional time-based maintenance strategies have proven inadequate for detecting early stage pump condition degradation, meaning corrective actions are generally taken only after a failure has occurred. This study aims to develop a predictive maintenance system based on vibration monitoring to support the implementation of condition-based maintenance (CBM). The system architecture utilizes an ADXL335 sensor for real-time vibration data acquisition, an ESP32 for data transmission, a Raspberry Pi as the data processing hub, and Node-RED as the monitoring interface. The vibration data undergoes preprocessing, time-domain statistical feature extraction, feature engineering, Health Index construction, and a clustering process that ultimately implements the Gaussian Mixture Model (GMM) as an optimization over the initial K-Means-based design. Furthermore, a reliability analysis is conducted using the Weibull distribution based on the temporal interval of precursor events. The results demonstrate that the system successfully classifies pump conditions into three main clusters with a Silhouette Score of 0.5948. The Weibull analysis yields a shape parameter (β) of 0.9865, a scale parameter (η) of 64.6112 hours, and a Mean Time Between Precursors (MTBP) of 64.99 hours. System validation achieves an Accuracy of 98.54%, a Recall of 100.00%, a Precision of 50.00%, an F1-score of 66.67%, and provides an Average Lead Time of 69.94 hours (≈ 3 days) prior to maintenance execution. These findings indicate that the system is highly capable of providing early warnings to support the formulation of more effective condition-based maintenance strategies in industrial environments.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Predictive Maintenance, Vibration Monitoring, Axial Piston Pump, Health Index, Gaussian Mixture Model, Weibull Analysis |
| Subjects: | Q Science T Technology > TJ Mechanical engineering and machinery T Technology > TJ Mechanical engineering and machinery > TJ174 Maintenance and repair of machinery T Technology > TJ Mechanical engineering and machinery > TJ217.6 Predictive Control T Technology > TJ Mechanical engineering and machinery > TJ910 Electric pumping machinery |
| Divisions: | Faculty of Vocational > 36304-Automation Electronic Engineering |
| Depositing User: | Wayan Indra Septian |
| Date Deposited: | 04 Aug 2026 04:18 |
| Last Modified: | 04 Aug 2026 04:18 |
| URI: | http://repository.its.ac.id/id/eprint/142868 |
Actions (login required)
 |
View Item |