Predicting crude unit failures and production impact using lagging maintenance indicators in oil refineries

Abstract Oil refineries operate in environments of high pressure and temperature. Therefore, maintaining the availability of these assets is a challenge to the maintenance teams. The paper deals with the development of maintenance key performance indicators (KPIs), considering both leading (post-failure) and lagging metrics for maintenance performance. Traditionally, these KPIs are looked upon in isolation; possible relationships between them are not considered. The present study investigated the relationships between various lagging maintenance KPIs relating production rate to equipment availability. The study was conducted under the refinery crude unit, which is generally considered one of the most important parts of any oil refinery. A regression model was developed that empirically relates equipment availability with the production rate. This was further extended to an empirical equation correlating equipment availability and downtime duration with various downtime factors. The results indicate that the availability of equipment has a high correlation with the production rate. This relationship is driven by factors of downtime, which have been sorted according to duration and type. These equations greatly support the management decisions in the development of maintenance strategies. In addition, the research quantifies the costs related to each factor of downtime, which can be applied in formulating future maintenance strategies.