Stochastic scheduling of workforce-constrained preventive maintenance activities in petroleum plants

Abstract

Maintenance is one of the major contributors of fixed cost or cost of ownership in a refinery or a chemical plant with a typical cost of $40–$70 million per annum. In order to reduce such huge cost, focus should be directed on the whole cycle of preventive maintenance (PM) which includes scheduling, wrench time (direct time working crew spends on maintenance), indirect time resulted due to different practical reasons such as work permits, weather disruptions, site availability, and corrective repairs. In this article, a stochastic modeling approach is proposed to schedule PM activities for an oil refinery considering both maintenance direct and indirect time uncertainties. The model is based on the minimization of total adjusted tardiness times and subject to limited amount of multiple-skill-level maintenance crew, specific time horizon and budget limit. The PM activities requires different crew skill set and prioritized based on expected cost due to risk of equipment failures when PM is delayed beyond its due date. The tardiness time is adjusted based on PM activity priority level. A real-life example where uncertainties are quantified using historical data will be used to demonstrate the validity of the proposed approach.