SISTEM MONITORING KONDISI PISAU DIE CUTTING DAN ANALISIS KEANDALAN (RELIABILITY ANALYSIS) DENGAN PENDEKATAN DISTRIBUSI WEIBULL DI PT.XYZ

Abstrak

Penelitian ini bertujuan untuk menganalisis keandalan pisau die cutting di PT. XYZ menggunakan distribusi Weibull serta merancang sistem monitoring kondisi untuk memprediksi waktu penggantian optimal. Latar belakang penelitian didasari oleh downtime tidak terjadwal akibat ketidakpastian waktu penggantian pisau dan belum adanya sistem prediktif di perusahaan. Data historis dari 30 pisau die cutting yang gagal selama Januari–April 2026 dianalisis melalui statistik deskriptif, uji distribusi, estimasi parameter Weibull dengan MLE, perhitungan fungsi keandalan R(t) dan hazard h(t), MTTF, median, serta penentuan waktu inspeksi dan penggantian optimal. Hasil penelitian menunjukkan data TTF memiliki variabilitas sangat tinggi (rentang 6.950–1.733.875 impression, mean 270.212, median 142.303). Distribusi Weibull 2 parameter terpilih sebagai model terbaik (p-value=0,042) dengan nilai shape β=0,671 dan scale η=185.742 impression, yang mengindikasikan pola infant mortality (β<1), bukan wear-out. MTTF diperoleh 239.050 impression dan median 107.730 impression. Waktu inspeksi ditetapkan pada 48.846 impression (keandalan 75%) dan penggantian optimal pada 142.303 impression (keandalan 50%). Sistem monitoring tiga zona warna (hijau, kuning, merah) dirancang untuk memandu teknisi dalam pengambilan keputusan penggantian pisau secara objektif. Kesimpulan penelitian ini adalah pola kegagalan pisau die cutting di PT. XYZ adalah infant mortality yang didominasi kerusakan rompang (63,3%). Model yang dibangun memiliki tingkat validitas baik dengan R²=0,942 dan error rata-rata 4,1%. PT. XYZ disarankan mengimplementasikan sistem monitoring tiga zona warna dengan batas inspeksi 48.846 impression dan penggantian optimal 142.303 impression, mengevaluasi kualitas pemasok pisau secara berkala, serta memperbarui model keandalan setiap 6 bulan. Penelitian ini berkontribusi pada pengembangan ilmu teknik industri khususnya penerapan analisis keandalan pada komponen die cutting di industri kemasan, serta memberikan solusi praktis bagi PT. XYZ dalam meningkatkan efektivitas perawatan dan produktivitas produksi

This study aims to analyze the reliability of die cutting blades at PT. XYZ using the Weibull distribution and to design a condition monitoring system for predicting optimal replacement time. The background of this research is based on unplanned downtime due to uncertainty in blade replacement and the absence of a predictive system in the company. Historical data from 30 failed die cutting blades during January–April 2026 were analyzed through descriptive statistics, goodness of fit test, Weibull parameter estimation using MLE, reliability function R(t) and hazard function h(t), MTTF, median, and determination of optimal inspection and replacement times. The results showed that TTF data had very high variability (range 6,950–1,733,875 impressions, mean 270,212, median 142,303). The 2-parameter Weibull distribution was selected as the best model (p-value=0.042) with shape parameter β=0.671 and scale parameter η=185,742 impressions, indicating an infant mortality pattern (β<1), not wear-out. MTTF obtained was 239,050 impressions and median was 107,730 impressions. Inspection time was set at 48,846 impressions (75% reliability) and optimal replacement at 142,303 impressions (50% reliability). A three-zone color monitoring system (green, yellow, red) was designed to guide technicians in making objective blade replacement decisions. The conclusion of this study is that the failure pattern of die cutting blades at PT. XYZ is infant mortality, dominated by chipped damage (63.3%). The developed model has good validity with R²=0.942 and average error of 4.1%. PT. XYZ is advised to implement a three-zone color monitoring system with inspection limit of 48,846 impressions and optimal replacement of 142,303 impressions, to evaluate blade supplier quality periodically, and to update the reliability model every 6 months. This research contributes to the development of industrial engineering knowledge, particularly in the application of reliability analysis to die cutting components in the packaging industry, and provides practical solutions for PT. XYZ in improving maintenance effectiveness and production productivity