Optimasi Penjadwalan Preventive Maintenance Komponen Boiler Menggunakan Genetic Algorithm dengan Strategi Grouping

Abstract

Operational reliability of Boilers in Steam Power Plants (PLTU) is often undermined by reactive and fragmented maintenance strategies that trigger frequent unplanned shutdowns and operational inefficiencies. This study aims to optimize preventive maintenance schedules for nine critical Boiler components in Units 1 and 2 of PLTU Nagan Raya using a Genetic Algorithm-based Grouping Maintenance strategy. Failure behaviour of components such as Cyclone Separators, Coal Feeders, and Fans is first modelled using Weibull and Normal distributions, selected through the Anderson–Darling goodness-of-fit test. An optimization model is then developed to determine a single global maintenance interval that maximizes operating time while satisfying a minimum system reliability constraint of 60%. Simulation of individual component scheduling yields highly non-synchronized intervals, ranging from 14 to 78 days, which would result in excessive shutdown frequency if implemented directly. Through Genetic Algorithm optimization, an optimal grouped maintenance interval of 10.42 days is obtained, driven primarily by the Primary Air Fan as the system bottleneck. Although this interval sacrifices part of the remaining life of more durable components, the grouped strategy significantly reduces the total number of shutdowns and simplifies maintenance planning. The main contribution of this study is a reliability-constrained grouped maintenance scheduling model for multi-component boiler systems in coal-fired power plants, providing a practical decision-support tool for improving unit availability.