Developing a Quantitative Risk-based Methodology for Maintenance Scheduling Using Bayesian Network

The main objective of the maintenance process is to increase equipment's life while maintaining the safety and reliability of the process systems. The maintenance planning concerns identification of what and how to inspect, how often to inspect, and what maintenance actions to be taken. Even though the maintenance may be used as an effective means for controlling the degradation of systems, the procedures may also have considerable impact on the operation. It results in direct and indirect economic consequences in terms of shutdowns and unavailability of systems. Therefore, it is necessary to plan maintenance such that a balance is achieved between the expected benefit and the corresponding economic consequences implied by these activities. The objective of this research is to integrate predictive and preventive maintenance strategies in an optimal way to maintain the desired availability and safety integrity level while minimizing the maintenance intervals. The outcome of this work would help to conserve resources while maintaining overall system availability and the safety. The results showed that the risk-based methodology developed using Bayesian Network increases the reliability of the equipment and also optimizes the cost of maintenance. Application of the developed methodology is demonstrated on the maintenance of a power plant as a case study.