Diesel Engine Generator (DEG) 300 kVA berfungsi sebagai sumber daya cadangan untuk memastikan kontinuitas operasional saat terjadi pemadaman listrik. Namun, dalam periode Mei–Agustus 2024, ditemukan kegagalan berulang pada mekanisme auto start, yang berpotensi mengganggu operasional fasilitas vital. Penelitian ini menggunakan pendekatan kuantitatif dengan menganalisis failure rate, reliability, mean time between failure, serta failure probability. Root Cause Analysis diterapkan sebagai metode sistematis untuk mengidentifikasi akar penyebab kegagalan. Hasil analisis menunjukkan failure probability tertinggi terjadi pada Juni (86,5%) dengan reliability terendah (13,5%). Rata-rata failure rate mencapai 2,54 kegagalan per jam, melebihi standar yang disyaratkan ≤0,02 kegagalan per jam. Performa terbaik terjadi pada Juli (failure probability 0%, reliability 100%), namun sistem kembali tidak stabil pada Agustus. Nilai mean time between failure terendah terjadi pada Juni, artinya sistem mengalami kegagalan dalam interval yang lebih singkat dibandingkan standar operasional. Root Cause Analysis mengidentifikasi bug firmware Real Time Clock dan kurangnya inspeksi berkala sebagai penyebab utama dan kurangnya redundansi Real Time Clock (RPN = 225) serta bug firmware Real Time Clock (RPN = 210) sebagai risiko utama. Untuk meningkatkan keandalan sistem, disarankan implementasi Real Time Clock cadangan, pembaruan firmware, serta perbaikan prosedur inspeksi dan troubleshooting agar mekanisme auto start lebih stabil dan dapat diandalkan.

The 300 kVA Diesel Engine Generator serves as a backup power source to ensure operational continuity during power outages. However, between May and August 2024, recurring failures in the auto-start mechanism were identified, potentially disrupting critical facility operations. This study employs a quantitative approach by analyzing failure rate, reliability, mean time between failure, and failure probability. Root Cause Analysis is applied to identify the underlying causes of failure. The results show that the highest failure probability occurred in June (86.5%) with the lowest reliability (13.5%). The average failure rate reached 2.54 failures per hour, exceeding the operational standard of ≤0.02 failures per hour. The best performance was in July (failure probability 0%, reliability 100%), but the system became unstabel again in August. The lowest mean time between failure in June indicates shorter failure intervals than the operational standard. Root Cause Analysis identified a Real Time Clock firmware bug and lack of regular inspections as primary causes, with insufficient redundancy (RPN = 225) and the firmware bug (RPN = 210) as major risks. To improve reliability, implementing a backup Real Time Clock, updating firmware, and optimizing inspections and troubleshooting are recommended.