A case study on the replacement policy for a pan system of sugar industry

In sugar production, the vacuum pans used for crystallisation constitute one of most important systems and replacement costs are a significant capital expenditure. The maintenance strategy for pans is currently conducted on a time-based approach. This paper proposes a condition-based maintenance (CBM) model for pans, which is applied to a real multi-component pan operating in an Australian sugar mill. The maintenance policy is formulated as a Markov Decision Process (MDP) and solved via approximate dynamic programming. The results show a 10% saving on total maintenance costs compared to the time-based maintenance strategy.